binding + tests

* use int64 stride everywhere

* fix ext

* fix ext

* more shape + cleanup

* one more

* few more

.pre-commit-config.yaml

@@ -1,13 +1,14 @@
 repos:
 -   repo: https://github.com/pre-commit/mirrors-clang-format
-    rev: v18.1.8
+    rev: v19.1.4
     hooks:
     -   id: clang-format
 # Using this mirror lets us use mypyc-compiled black, which is about 2x faster
 -   repo: https://github.com/psf/black-pre-commit-mirror
-    rev: 24.8.0
+    rev: 24.10.0
     hooks:
     -   id: black
+    
 -   repo: https://github.com/pycqa/isort
     rev: 5.13.2
     hooks:

CMakeLists.txt

@@ -24,7 +24,7 @@ option(MLX_METAL_JIT "Use JIT compilation for Metal kernels" OFF)
 option(BUILD_SHARED_LIBS "Build mlx as a shared library" OFF)
 
 if(NOT MLX_VERSION)
-  set(MLX_VERSION 0.19.3)
+  set(MLX_VERSION 0.21.1)
 endif()
 
 # --------------------- Processor tests -------------------------
@@ -34,8 +34,6 @@ message(
     "Building MLX for ${CMAKE_SYSTEM_PROCESSOR} processor on ${CMAKE_SYSTEM_NAME}"
 )
 
-set(MLX_BUILD_ARM OFF)
-
 if(${CMAKE_SYSTEM_NAME} MATCHES "Darwin")
   if(${CMAKE_SYSTEM_PROCESSOR} MATCHES "x86_64")
     if(NOT MLX_ENABLE_X64_MAC)
@@ -57,10 +55,6 @@ else()
   message(WARNING "MLX is prioritised for Apple silicon systems using macOS.")
 endif()
 
-if(${CMAKE_SYSTEM_PROCESSOR} MATCHES "arm64")
-  set(MLX_BUILD_ARM ON)
-endif()
-
 # ----------------------------- Lib -----------------------------
 
 include(FetchContent)
@@ -89,25 +83,27 @@ elseif(MLX_BUILD_METAL)
   # Throw an error if xcrun not found
   execute_process(
     COMMAND zsh "-c" "/usr/bin/xcrun -sdk macosx --show-sdk-version"
-    OUTPUT_VARIABLE MACOS_VERSION COMMAND_ERROR_IS_FATAL ANY)
+    OUTPUT_VARIABLE MACOS_SDK_VERSION COMMAND_ERROR_IS_FATAL ANY)
 
-  if(${MACOS_VERSION} LESS 14.0)
+  if(${MACOS_SDK_VERSION} LESS 14.0)
     message(
       FATAL_ERROR
         "MLX requires macOS SDK >= 14.0 to be built with MLX_BUILD_METAL=ON")
   endif()
-  message(STATUS "Building with SDK for macOS version ${MACOS_VERSION}")
+  message(STATUS "Building with macOS SDK version ${MACOS_SDK_VERSION}")
 
   set(METAL_CPP_URL
       https://developer.apple.com/metal/cpp/files/metal-cpp_macOS15_iOS18-beta.zip
   )
-  # Get the metal version
+
+  if(NOT CMAKE_OSX_DEPLOYMENT_TARGET STREQUAL "")
+    set(XCRUN_FLAGS "-mmacosx-version-min=${CMAKE_OSX_DEPLOYMENT_TARGET}")
+  endif()
   execute_process(
     COMMAND
       zsh "-c"
-      "echo \"__METAL_VERSION__\" | xcrun -sdk macosx metal -E -x metal -P - | tail -1 | tr -d '\n'"
+      "echo \"__METAL_VERSION__\" | xcrun -sdk macosx metal ${XCRUN_FLAGS} -E -x metal -P - | tail -1 | tr -d '\n'"
     OUTPUT_VARIABLE MLX_METAL_VERSION COMMAND_ERROR_IS_FATAL ANY)
-
   FetchContent_Declare(metal_cpp URL ${METAL_CPP_URL})
 
   FetchContent_MakeAvailable(metal_cpp)
@@ -115,13 +111,11 @@ elseif(MLX_BUILD_METAL)
     mlx PUBLIC $<BUILD_INTERFACE:${metal_cpp_SOURCE_DIR}>
                $<INSTALL_INTERFACE:include/metal_cpp>)
   target_link_libraries(mlx PUBLIC ${METAL_LIB} ${FOUNDATION_LIB} ${QUARTZ_LIB})
-
-  add_compile_definitions("MLX_METAL_VERSION=${MLX_METAL_VERSION}")
 endif()
 
 if(MLX_BUILD_CPU)
   find_library(ACCELERATE_LIBRARY Accelerate)
-  if(MLX_BUILD_ARM AND ACCELERATE_LIBRARY)
+  if(ACCELERATE_LIBRARY)
     message(STATUS "Accelerate found ${ACCELERATE_LIBRARY}")
     set(MLX_BUILD_ACCELERATE ON)
     target_link_libraries(mlx PUBLIC ${ACCELERATE_LIBRARY})
@@ -160,6 +154,13 @@ if(MLX_BUILD_CPU)
     message(STATUS "Blas include " ${BLAS_INCLUDE_DIRS})
     target_include_directories(mlx PRIVATE ${BLAS_INCLUDE_DIRS})
     target_link_libraries(mlx PUBLIC ${BLAS_LIBRARIES})
+
+    if(WIN32)
+      find_package(dlfcn-win32 REQUIRED)
+      message(STATUS "dlfcn-win32 lib " ${dlfcn-win32_LIBRARIES})
+      message(STATUS "dlfcn-win32 include " ${dlfcn-win32_INCLUDE_DIRS})
+      target_link_libraries(mlx PUBLIC ${dlfcn-win32_LIBRARIES})
+    endif()
   endif()
 else()
   set(MLX_BUILD_ACCELERATE OFF)

benchmarks/python/sdpa_bench.py

@@ -1,62 +1,189 @@
+# Copyright © 2024 Apple Inc.
+
 import argparse
 import math
+import os
+import subprocess
+import time
 
 import mlx.core as mx
-from time_utils import time_fn
+import numpy as np
 
-MAX_SEQ = 300
+device_name = subprocess.check_output(["sysctl", "-n", "machdep.cpu.brand_string"])
-START_SEQ = 100
+device_name = device_name.decode("utf-8").strip("\n")
-SEQ_INCREMENT = 50
+
+N_warmup = 5
+N_iter_bench = 40
+N_iter_func = 8
 
 
-def time_self_attention_primitives():
+def bench(f, *args):
-    mx.random.seed(3)
+    for i in range(N_warmup):
-    B = 2
+        f(*args)
-    H = 38
-    D = 64
-    for R in range(START_SEQ, MAX_SEQ, SEQ_INCREMENT):
-        q = mx.random.uniform(shape=(B, H, R, D))
-        k = mx.random.uniform(shape=(B, H, R, D))
-        v = mx.random.uniform(shape=(B, H, R, D))
-        scale = 1.0 / math.sqrt(float(D))
-        mx.eval(q, k, v)
 
-        def sdpa_primitives(qs, ks, vs, alpha):
+    s = time.perf_counter_ns()
-            s = (alpha * qs) @ ks.transpose(0, 1, 3, 2)
+    for i in range(N_iter_bench):
-            p = mx.softmax(s.astype(mx.float32), axis=-1).astype(s.dtype)
+        f(*args)
-            o = p @ vs
+    e = time.perf_counter_ns()
-            return o
+    return (e - s) * 1e-9
-
-        time_fn(sdpa_primitives, q, k, v, scale)
 
 
-def time_self_attention_sdpa():
+def mlx_sdpa_fused_inner(q, k, v, scale):
-    mx.random.seed(3)
+    return mx.fast.scaled_dot_product_attention(q, k, v, scale=scale, mask=None)
-    B = 2
-    H = 38
-    D = 64
-    for R in range(START_SEQ, MAX_SEQ, SEQ_INCREMENT):
-        q = mx.random.uniform(shape=(B, H, R, D))
-        k = mx.random.uniform(shape=(B, H, R, D))
-        v = mx.random.uniform(shape=(B, H, R, D))
-        scale = 1.0 / math.sqrt(float(D))
-        mx.eval(q, k, v)
 
-        def sdpa_fused(qs, ks, vs, alpha):
-            o = mx.fast.scaled_dot_product_attention(qs, ks, vs, scale=alpha)
-            return o
 
-        time_fn(sdpa_fused, q, k, v, scale)
+def mlx_sdpa_unfused_inner(q, k, v, scale, f32softmax=False):
+    q_dtype = q.dtype
+    q = q * mx.array(scale, q_dtype)
+    n_q_heads = q.shape[-3]
+    n_kv_heads = k.shape[-3]
+    n_repeats = n_q_heads // n_kv_heads
+
+    B = q.shape[0]
+    L = q.shape[2]
+
+    if n_repeats > 1:
+        q = mx.reshape(q, [B, n_kv_heads, n_repeats, L, -1])
+        k = mx.expand_dims(k, 2)
+        v = mx.expand_dims(v, 2)
+
+    scores = q @ mx.swapaxes(k, -1, -2)
+    if f32softmax:
+        scores = mx.softmax(scores.astype(mx.float32), axis=-1).astype(q_dtype)
+    else:
+        scores = mx.softmax(scores, axis=-1)
+
+    out = scores @ v
+    if n_repeats > 1:
+        out = mx.reshape(out, [B, n_q_heads, L, -1])
+
+    return out
+
+
+def mlx_spda_unfused(q, k, v, scale, transpose):
+    q_out = q
+    if transpose:
+        k = mx.transpose(k, (0, 2, 1, 3))
+        v = mx.transpose(v, (0, 2, 1, 3))
+
+    for i in range(N_iter_func):
+        if transpose:
+            q_out = mx.transpose(q_out, (0, 2, 1, 3))
+        q_out = mlx_sdpa_unfused_inner(q_out, k, v, scale)
+        if transpose:
+            q_out = mx.transpose(q_out, (0, 2, 1, 3))
+
+    mx.eval(q_out)
+    return q_out
+
+
+def mlx_spda_fused(q, k, v, scale, transpose):
+    q_out = q
+    if transpose:
+        k = mx.transpose(k, (0, 2, 1, 3))
+        v = mx.transpose(v, (0, 2, 1, 3))
+
+    for i in range(N_iter_func):
+        if transpose:
+            q_out = mx.transpose(q_out, (0, 2, 1, 3))
+        q_out = mlx_sdpa_fused_inner(q_out, k, v, scale)
+        if transpose:
+            q_out = mx.transpose(q_out, (0, 2, 1, 3))
+
+    mx.eval(q_out)
+    return q_out
+
+
+def bench_shape(B, qsl, ksl, head_dim, n_q_heads, n_kv_heads, np_dtype, transpose=True):
+    shape_q = (
+        (B, qsl, n_q_heads, head_dim) if transpose else (B, n_q_heads, qsl, head_dim)
+    )
+    shape_kv = (
+        (B, ksl, n_kv_heads, head_dim) if transpose else (B, n_kv_heads, ksl, head_dim)
+    )
+
+    q_np = np.random.normal(0.0, 1.0 / math.sqrt(head_dim), shape_q).astype(np_dtype)
+    k_np = np.random.normal(0.0, 1.0 / math.sqrt(head_dim), shape_kv).astype(np_dtype)
+    v_np = np.random.normal(0.0, 1.0 / math.sqrt(head_dim), shape_kv).astype(np_dtype)
+
+    scale = math.sqrt(1.0 / head_dim)
+
+    q_mx = mx.array(q_np)
+    k_mx = mx.array(k_np)
+    v_mx = mx.array(v_np)
+
+    time_mlx_unfused = bench(mlx_spda_unfused, q_mx, k_mx, v_mx, scale, transpose)
+    time_mlx_fused = bench(mlx_spda_fused, q_mx, k_mx, v_mx, scale, transpose)
+
+    if transpose:
+        q_mx = mx.transpose(q_mx, (0, 2, 1, 3))
+        k_mx = mx.transpose(k_mx, (0, 2, 1, 3))
+        v_mx = mx.transpose(v_mx, (0, 2, 1, 3))
+
+    o_mlx_fused = mlx_sdpa_fused_inner(q_mx, k_mx, v_mx, scale)
+    o_mlx_unfused = mlx_sdpa_unfused_inner(q_mx, k_mx, v_mx, scale, f32softmax=True)
+
+    atol = 1e-5 if np_dtype == np.float32 else 1e-4
+
+    if not mx.allclose(o_mlx_fused, o_mlx_unfused, atol=atol):
+        print(
+            f"Failed at (B: {B}, qsl: {qsl}, ksl: {ksl}, head_dim: {head_dim}, n_qh: {n_q_heads}, n_kvh: {n_kv_heads}) [tpose = {transpose}] with max(|a - b|) = {mx.max(mx.abs(o_mlx_unfused - o_mlx_fused)):3.2e}"
+        )
+
+    return time_mlx_fused, time_mlx_unfused
+
+
+def get_gflop_count(B, M, N, K):
+    return float(2.0 * N_iter_bench * N_iter_func * B * M * N * K) / float(1024.0**3)
 
 
 if __name__ == "__main__":
-    parser = argparse.ArgumentParser("MLX benchmarks.")
+    parser = argparse.ArgumentParser(description="Run gemm benchmarks")
-    parser.add_argument("--gpu", action="store_true", help="Use the Metal back-end.")
-    args = parser.parse_args()
-    if args.gpu:
-        mx.set_default_device(mx.gpu)
-    else:
-        mx.set_default_device(mx.cpu)
 
-    time_self_attention_sdpa()
+    dtypes = ("float16", "float32")[:1]
-    time_self_attention_primitives()
+    transposes = (False,)
+
+    # fmt: off
+    shapes_64 = (
+        # (  B,   qsl,   ksl, head_dim, n_qh, n_kvh)
+          (  1,    32,    32,       64,   32,    32),
+          (  1,    64,    64,       64,   32,    32),
+          (  1,   128,   128,       64,   32,    32),
+          (  1,   256,   256,       64,   32,    32),
+          (  1,   512,   512,       64,   32,    32),
+          (  1,  1024,  1024,       64,   32,    32),
+          (  1,  2048,  2048,       64,   32,    32),
+          (  1,  4096,  4096,       64,   32,    32),
+    )
+
+    shapes_80 = (
+        # (  B,   qsl,   ksl, head_dim, n_qh, n_kvh)
+          (  1,  1024,  1024,       80,   32,    32),
+          (  1,  2048,  2048,       80,   32,    32),
+          (  1,  4096,  4096,       80,   32,    32),
+    )
+
+    shapes_128 = (
+        # (  B,   qsl,   ksl, head_dim, n_qh, n_kvh)
+          (  1,  1024,  1024,      128,   32,    32),
+          (  1,  2048,  2048,      128,   32,    32),
+          (  1,  4096,  4096,      128,   32,    32),
+    )
+    # fmt: on
+
+    shapes = shapes_64 + shapes_80 + shapes_128
+
+    print("  B,   qsl,   ksl, hdim, n_qh, n_kvh, tpose,   dtype, t_unfs, t_fuse, diff%")
+
+    for dtype in dtypes:
+        for transpose in transposes:
+            for B, qsl, ksl, head_dim, n_q_heads, n_kv_heads in shapes:
+                np_dtype = getattr(np, dtype)
+                time_mlx_fused, time_mlx_unfused = bench_shape(
+                    B, qsl, ksl, head_dim, n_q_heads, n_kv_heads, np_dtype, transpose
+                )
+                diff = time_mlx_unfused / time_mlx_fused - 1.0
+                t_str = 1 if transpose else 0
+                print(
+                    f"{B:3d}, {qsl:5d}, {ksl:5d}, {head_dim:4d}, {n_q_heads:4d}, {n_kv_heads:5d}, {t_str:5d}, {dtype}, {time_mlx_unfused: 2.3f}, {time_mlx_fused: 2.3f}, {100. * diff:+5.2f}%"
+                )

benchmarks/python/sdpa_vector_bench.py

@@ -4,42 +4,51 @@ import math
 import mlx.core as mx
 from time_utils import time_fn
 
-L = 1024
+L = 16384
 H = 32
-H_k = 32 // 4
+H_k = H // 4
 D = 128
+dtype = mx.float16
+loops = 10
 
 
 def attention(q, k, v):
-    B, Hq, L, D = q.shape
+    def _sdpa(q, k, v):
-    _, Hk, S, _ = k.shape
+        B, Hq, L, D = q.shape
-    q = q.reshape(B, Hk, Hq // Hk, L, D)
+        _, Hk, S, _ = k.shape
-    k = k[:, :, None, :, :]
+        q = q.reshape(B, Hk, Hq // Hk, L, D)
-    v = v[:, :, None, :, :]
+        k = k[:, :, None, :, :]
-    s = q @ k.transpose(0, 1, 2, 4, 3)
+        v = v[:, :, None, :, :]
-    p = mx.softmax(s.astype(mx.float32), axis=-1).astype(s.dtype)
+        s = q @ k.transpose(0, 1, 2, 4, 3)
-    o = p @ v
+        p = mx.softmax(s.astype(mx.float32), axis=-1).astype(s.dtype)
-    return o.reshape(B, Hq, L, D)
+        o = p @ v
+        return o.reshape(B, Hq, L, D)
+
+    for i in range(loops):
+        q = _sdpa(q, k, v)
+    return q
 
 
 def sdpa(q, k, v):
-    return mx.fast.scaled_dot_product_attention(q, k, v, scale=1.0)
+    for i in range(loops):
+        q = mx.fast.scaled_dot_product_attention(q, k, v, scale=1.0)
+    return q
 
 
 def time_self_attention_primitives():
     mx.random.seed(3)
-    q = mx.random.uniform(shape=(1, H, 1, D))
+    q = mx.random.uniform(shape=(1, H, 1, D)).astype(dtype)
-    k = mx.random.uniform(shape=(1, H_k, L, D))
+    k = mx.random.uniform(shape=(1, H_k, L, D)).astype(dtype)
-    v = mx.random.uniform(shape=(1, H_k, L, D))
+    v = mx.random.uniform(shape=(1, H_k, L, D)).astype(dtype)
     mx.eval(q, k, v)
     time_fn(attention, q, k, v)
 
 
 def time_self_attention_sdpa():
     mx.random.seed(3)
-    q = mx.random.uniform(shape=(1, H, 1, D))
+    q = mx.random.uniform(shape=(1, H, 1, D)).astype(dtype)
-    k = mx.random.uniform(shape=(1, H_k, L, D))
+    k = mx.random.uniform(shape=(1, H_k, L, D)).astype(dtype)
-    v = mx.random.uniform(shape=(1, H_k, L, D))
+    v = mx.random.uniform(shape=(1, H_k, L, D)).astype(dtype)
     mx.eval(q, k, v)
     time_fn(sdpa, q, k, v)
 

docs/src/dev/extensions.rst

@@ -420,8 +420,8 @@ element in the output.
             constant const float& alpha [[buffer(3)]],
             constant const float& beta [[buffer(4)]],
             constant const int* shape [[buffer(5)]],
-            constant const size_t* x_strides [[buffer(6)]],
+            constant const int64_t* x_strides [[buffer(6)]],
-            constant const size_t* y_strides [[buffer(7)]],
+            constant const int64_t* y_strides [[buffer(7)]],
             constant const int& ndim [[buffer(8)]],
             uint index [[thread_position_in_grid]]) {
         // Convert linear indices to offsets in array
@@ -438,24 +438,10 @@ each instantiation a unique host name so we can identify it.
 
 .. code-block:: C++
 
-    #define instantiate_axpby(type_name, type)              \
+    instantiate_kernel("axpby_general_float32", axpby_general, float)
-        template [[host_name("axpby_general_" #type_name)]] \
+    instantiate_kernel("axpby_general_float16", axpby_general, float16_t)
-        [[kernel]] void axpby_general<type>(                \
+    instantiate_kernel("axpby_general_bfloat16", axpby_general, bfloat16_t)
-            device const type* x [[buffer(0)]],             \
+    instantiate_kernel("axpby_general_complex64", axpby_general, complex64_t)
-            device const type* y [[buffer(1)]],             \
-            device type* out [[buffer(2)]],                 \
-            constant const float& alpha [[buffer(3)]],      \
-            constant const float& beta [[buffer(4)]],       \
-            constant const int* shape [[buffer(5)]],        \
-            constant const size_t* x_strides [[buffer(6)]], \
-            constant const size_t* y_strides [[buffer(7)]], \
-            constant const int& ndim [[buffer(8)]],         \
-            uint index [[thread_position_in_grid]]);
-
-    instantiate_axpby(float32, float);
-    instantiate_axpby(float16, half);
-    instantiate_axpby(bfloat16, bfloat16_t);
-    instantiate_axpby(complex64, complex64_t);
 
 The logic to determine the kernel, set the inputs, resolve the grid dimensions,
 and dispatch to the GPU are contained in :meth:`Axpby::eval_gpu` as shown
@@ -494,7 +480,7 @@ below.
 
         // Prepare to encode kernel
         auto& compute_encoder = d.get_command_encoder(s.index);
-        compute_encoder->setComputePipelineState(kernel);
+        compute_encoder.set_compute_pipeline_state(kernel);
 
         // Kernel parameters are registered with buffer indices corresponding to
         // those in the kernel declaration at axpby.metal
@@ -509,14 +495,14 @@ below.
         compute_encoder.set_output_array(out, 2);
 
         // Encode alpha and beta
-        compute_encoder->setBytes(&alpha_, sizeof(float), 3);
+        compute_encoder.set_bytes(alpha_, 3);
-        compute_encoder->setBytes(&beta_, sizeof(float), 4);
+        compute_encoder.set_bytes(beta_, 4);
 
         // Encode shape, strides and ndim
-        compute_encoder->setBytes(x.shape().data(), ndim * sizeof(int), 5);
+        compute_encoder.set_vector_bytes(x.shape(), 5);
-        compute_encoder->setBytes(x.strides().data(), ndim * sizeof(size_t), 6);
+        compute_encoder.set_vector_bytes(x.strides(), 6);
-        compute_encoder->setBytes(y.strides().data(), ndim * sizeof(size_t), 7);
+        compute_encoder.set_bytes(y.strides(), 7);
-        compute_encoder->setBytes(&ndim, sizeof(int), 8);
+        compute_encoder.set_bytes(ndim, 8);
 
         // We launch 1 thread for each input and make sure that the number of
         // threads in any given threadgroup is not higher than the max allowed
@@ -530,7 +516,7 @@ below.
 
         // Launch the grid with the given number of threads divided among
         // the given threadgroups
-        compute_encoder.dispatchThreads(grid_dims, group_dims);
+        compute_encoder.dispatch_threads(grid_dims, group_dims);
     }
 
 We can now call the :meth:`axpby` operation on both the CPU and the GPU!

docs/src/install.rst

@@ -209,7 +209,7 @@ Metal library by run-time compiling kernels the first time they are used in MLX
 on a given machine. Note run-time compilation incurs a cold-start cost which can
 be anwywhere from a few hundred millisecond to a few seconds depending on the
 application. Once a kernel is compiled, it will be cached by the system. The
-Metal kernel cache persists accross reboots.
+Metal kernel cache persists across reboots.
 
 Troubleshooting
 ^^^^^^^^^^^^^^^

docs/src/python/fast.rst

@@ -12,5 +12,4 @@ Fast
   layer_norm
   rope
   scaled_dot_product_attention
-  affine_quantize
   metal_kernel

docs/src/python/nn/layers.rst

@@ -12,6 +12,7 @@ Layers
    ALiBi
    AvgPool1d
    AvgPool2d
+   AvgPool3d
    BatchNorm
    CELU
    Conv1d
@@ -41,6 +42,7 @@ Layers
    LSTM
    MaxPool1d
    MaxPool2d
+   MaxPool3d
    Mish
    MultiHeadAttention
    PReLU

docs/src/usage/function_transforms.rst

@@ -184,8 +184,8 @@ Let's time these two different versions:
   print(timeit.timeit(lambda: mx.eval(naive_add(xs, ys)), number=100))
   print(timeit.timeit(lambda: mx.eval(vmap_add(xs, ys)), number=100))
 
-On an M1 Max the naive version takes in total ``0.390`` seconds whereas the
+On an M1 Max the naive version takes in total ``5.639`` seconds whereas the
-vectorized version takes only ``0.025`` seconds, more than ten times faster.
+vectorized version takes only ``0.024`` seconds, more than 200 times faster.
 
 Of course, this operation is quite contrived. A better approach is to simply do
 ``xs + ys.T``, but for more complex functions :func:`vmap` can be quite handy.

examples/extensions/axpby/axpby.cpp

@@ -257,7 +257,7 @@ void Axpby::eval_gpu(
 
   // Prepare to encode kernel
   auto& compute_encoder = d.get_command_encoder(s.index);
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
 
   // Kernel parameters are registered with buffer indices corresponding to
   // those in the kernel declaration at axpby.metal
@@ -272,15 +272,15 @@ void Axpby::eval_gpu(
   compute_encoder.set_output_array(out, 2);
 
   // Encode alpha and beta
-  compute_encoder->setBytes(&alpha_, sizeof(float), 3);
+  compute_encoder.set_bytes(alpha_, 3);
-  compute_encoder->setBytes(&beta_, sizeof(float), 4);
+  compute_encoder.set_bytes(beta_, 4);
 
   // Encode shape, strides and ndim if needed
   if (!contiguous_kernel) {
-    compute_encoder->setBytes(x.shape().data(), ndim * sizeof(int), 5);
+    compute_encoder.set_vector_bytes(x.shape(), 5);
-    compute_encoder->setBytes(x.strides().data(), ndim * sizeof(size_t), 6);
+    compute_encoder.set_vector_bytes(x.strides(), 6);
-    compute_encoder->setBytes(y.strides().data(), ndim * sizeof(size_t), 7);
+    compute_encoder.set_bytes(y.strides(), 7);
-    compute_encoder->setBytes(&ndim, sizeof(int), 8);
+    compute_encoder.set_bytes(ndim, 8);
   }
 
   // We launch 1 thread for each input and make sure that the number of
@@ -295,7 +295,7 @@ void Axpby::eval_gpu(
 
   // Launch the grid with the given number of threads divided among
   // the given threadgroups
-  compute_encoder.dispatchThreads(grid_dims, group_dims);
+  compute_encoder.dispatch_threads(grid_dims, group_dims);
 }
 
 #else // Metal is not available

examples/extensions/axpby/axpby.metal

@@ -2,7 +2,6 @@
 
 #include <metal_stdlib>
 
-#include "mlx/backend/metal/kernels/bf16.h"
 #include "mlx/backend/metal/kernels/utils.h"
 
 template <typename T>
@@ -13,8 +12,8 @@ template <typename T>
     constant const float& alpha [[buffer(3)]],
     constant const float& beta [[buffer(4)]],
     constant const int* shape [[buffer(5)]],
-    constant const size_t* x_strides [[buffer(6)]],
+    constant const int64_t* x_strides [[buffer(6)]],
-    constant const size_t* y_strides [[buffer(7)]],
+    constant const int64_t* y_strides [[buffer(7)]],
     constant const int& ndim [[buffer(8)]],
     uint index [[thread_position_in_grid]]) {
   auto x_offset = elem_to_loc(index, shape, x_strides, ndim);
@@ -35,29 +34,14 @@ template <typename T>
       static_cast<T>(alpha) * x[index] + static_cast<T>(beta) * y[index];
 }
 
-#define instantiate_axpby(type_name, type)                               \
+// clang-format off
-  template [[host_name("axpby_general_" #type_name)]] [[kernel]] void    \
+#define instantiate_axpby(type_name, type)                             \
-  axpby_general<type>(                                                   \
+  instantiate_kernel("axpby_general_" #type_name, axpby_general, type) \
-      device const type* x [[buffer(0)]],                                \
+  instantiate_kernel(                                                  \
-      device const type* y [[buffer(1)]],                                \
+          "axpby_contiguous_" #type_name, axpby_contiguous, type)
-      device type* out [[buffer(2)]],                                    \
-      constant const float& alpha [[buffer(3)]],                         \
-      constant const float& beta [[buffer(4)]],                          \
-      constant const int* shape [[buffer(5)]],                           \
-      constant const size_t* x_strides [[buffer(6)]],                    \
-      constant const size_t* y_strides [[buffer(7)]],                    \
-      constant const int& ndim [[buffer(8)]],                            \
-      uint index [[thread_position_in_grid]]);                           \
-  template [[host_name("axpby_contiguous_" #type_name)]] [[kernel]] void \
-  axpby_contiguous<type>(                                                \
-      device const type* x [[buffer(0)]],                                \
-      device const type* y [[buffer(1)]],                                \
-      device type* out [[buffer(2)]],                                    \
-      constant const float& alpha [[buffer(3)]],                         \
-      constant const float& beta [[buffer(4)]],                          \
-      uint index [[thread_position_in_grid]]);
 
 instantiate_axpby(float32, float);
 instantiate_axpby(float16, half);
 instantiate_axpby(bfloat16, bfloat16_t);
-instantiate_axpby(complex64, complex64_t);
+instantiate_axpby(complex64, complex64_t);
+// clang-format on

examples/extensions/requirements.txt

@@ -1,4 +1,4 @@
 setuptools>=42
 cmake>=3.24
-mlx>=0.18.1
+mlx>=0.21.0
 nanobind==2.2.0

mlx.pc.in

@@ -28,10 +28,19 @@ endif()
 if (@MLX_BUILD_METAL@)
     set(MLX_BUILD_METAL @MLX_BUILD_METAL@)
     set(MLX_CXX_FLAGS ${MLX_CXX_FLAGS} -D_METAL_)
-    set_and_check(MLX_INCLUDE_DIRS 
+    set(MLX_INCLUDE_DIRS 
-        ${MLX_INCLUDE_DIRS} 
+        "${MLX_INCLUDE_DIRS};"
         @PACKAGE_CMAKE_INSTALL_INCLUDEDIR@/metal_cpp
     )
+    if(@MLX_METAL_VERSION@ GREATER_EQUAL 310)
+      set(MLX_INCLUDE_DIRS
+        "${MLX_INCLUDE_DIRS};"
+        @PACKAGE_CMAKE_INSTALL_INCLUDEDIR@/mlx/backend/metal/kernels/metal_3_1)
+    else()
+      set(MLX_INCLUDE_DIRS
+        "${MLX_INCLUDE_DIRS};"
+        @PACKAGE_CMAKE_INSTALL_INCLUDEDIR@/mlx/backend/metal/kernels/metal_3_0)
+    endif()
 endif()
 
 set_target_properties(mlx PROPERTIES
@@ -40,4 +49,4 @@ set_target_properties(mlx PROPERTIES
 )
 
 include(FindPackageHandleStandardArgs)
-find_package_handle_standard_args(MLX DEFAULT_MSG MLX_LIBRARY MLX_INCLUDE_DIRS)
+find_package_handle_standard_args(MLX DEFAULT_MSG MLX_LIBRARY MLX_INCLUDE_DIRS)

mlx/allocator.cpp

@@ -19,7 +19,7 @@ Buffer malloc(size_t size) {
 }
 
 void free(Buffer buffer) {
-  return allocator().free(buffer);
+  allocator().free(buffer);
 }
 
 Buffer CommonAllocator::malloc(size_t size, bool) {

mlx/array.cpp

@@ -1,5 +1,6 @@
 // Copyright © 2023-2024 Apple Inc.
 #include <functional>
+#include <unordered_map>
 
 #include "mlx/array.h"
 #include "mlx/ops.h"
@@ -30,7 +31,7 @@ array::array(const std::complex<float>& val, Dtype dtype /* = complex64 */)
 }
 
 array::array(
-    std::vector<int> shape,
+    Shape shape,
     Dtype dtype,
     std::shared_ptr<Primitive> primitive,
     std::vector<array> inputs)
@@ -41,7 +42,7 @@ array::array(
           std::move(inputs))) {}
 
 std::vector<array> array::make_arrays(
-    std::vector<std::vector<int>> shapes,
+    std::vector<Shape> shapes,
     const std::vector<Dtype>& dtypes,
     const std::shared_ptr<Primitive>& primitive,
     const std::vector<array>& inputs) {
@@ -73,11 +74,7 @@ array::array(std::initializer_list<int> data, Dtype dtype)
 }
 
 /* Build an array from a shared buffer */
-array::array(
+array::array(allocator::Buffer data, Shape shape, Dtype dtype, Deleter deleter)
-    allocator::Buffer data,
-    std::vector<int> shape,
-    Dtype dtype,
-    deleter_t deleter)
     : array_desc_(std::make_shared<ArrayDesc>(std::move(shape), dtype)) {
   set_data(data, deleter);
 }
@@ -125,7 +122,7 @@ bool array::is_tracer() const {
   return array_desc_->is_tracer && in_tracing() || retain_graph();
 }
 
-void array::set_data(allocator::Buffer buffer, deleter_t d) {
+void array::set_data(allocator::Buffer buffer, Deleter d) {
   array_desc_->data = std::make_shared<Data>(buffer, d);
   array_desc_->data_ptr = buffer.raw_ptr();
   array_desc_->data_size = size();
@@ -138,9 +135,9 @@ void array::set_data(allocator::Buffer buffer, deleter_t d) {
 void array::set_data(
     allocator::Buffer buffer,
     size_t data_size,
-    std::vector<size_t> strides,
+    Strides strides,
     Flags flags,
-    deleter_t d) {
+    Deleter d) {
   array_desc_->data = std::make_shared<Data>(buffer, d);
   array_desc_->data_ptr = buffer.raw_ptr();
   array_desc_->data_size = data_size;
@@ -150,7 +147,7 @@ void array::set_data(
 
 void array::copy_shared_buffer(
     const array& other,
-    const std::vector<size_t>& strides,
+    const Strides& strides,
     Flags flags,
     size_t data_size,
     size_t offset /* = 0 */) {
@@ -169,7 +166,7 @@ void array::copy_shared_buffer(const array& other) {
 
 void array::move_shared_buffer(
     array other,
-    const std::vector<size_t>& strides,
+    const Strides& strides,
     Flags flags,
     size_t data_size,
     size_t offset /* = 0 */) {
@@ -214,6 +211,8 @@ array::~array() {
     if (do_detach) {
       for (auto& s : siblings()) {
         for (auto& ss : s.siblings()) {
+          // Set to null here to avoid descending into array destructor
+          // for siblings
           ss.array_desc_ = nullptr;
         }
         s.array_desc_->siblings.clear();
@@ -234,13 +233,13 @@ void array::ArrayDesc::init() {
   }
 }
 
-array::ArrayDesc::ArrayDesc(std::vector<int> shape, Dtype dtype)
+array::ArrayDesc::ArrayDesc(Shape shape, Dtype dtype)
     : shape(std::move(shape)), dtype(dtype), status(Status::available) {
   init();
 }
 
 array::ArrayDesc::ArrayDesc(
-    std::vector<int> shape,
+    Shape shape,
     Dtype dtype,
     std::shared_ptr<Primitive> primitive,
     std::vector<array> inputs)
@@ -292,6 +291,14 @@ array::ArrayDesc::~ArrayDesc() {
     auto top = std::move(for_deletion.back());
     for_deletion.pop_back();
     append_deletable_inputs(*top);
+
+    // Clear out possible siblings to break circular references
+    for (auto& s : top->siblings) {
+      // Set to null here to avoid descending into top-level
+      // array destructor for siblings
+      s.array_desc_ = nullptr;
+    }
+    top->siblings.clear();
   }
 }
 

mlx/array.h

@@ -15,7 +15,10 @@ namespace mlx::core {
 
 // Forward declaration
 class Primitive;
-using deleter_t = std::function<void(allocator::Buffer)>;
+
+using Deleter = std::function<void(allocator::Buffer)>;
+using Shape = std::vector<int32_t>;
+using Strides = std::vector<int64_t>;
 
 class array {
   /* An array is really a node in a graph. It contains a shared ArrayDesc
@@ -33,7 +36,7 @@ class array {
   template <typename It>
   array(
       It data,
-      std::vector<int> shape,
+      Shape shape,
       Dtype dtype =
           TypeToDtype<typename std::iterator_traits<It>::value_type>());
 
@@ -49,15 +52,15 @@ class array {
   template <typename T>
   array(
       std::initializer_list<T> data,
-      std::vector<int> shape,
+      Shape shape,
       Dtype dtype = TypeToDtype<T>());
 
   /* Build an array from a buffer */
   array(
       allocator::Buffer data,
-      std::vector<int> shape,
+      Shape shape,
       Dtype dtype,
-      deleter_t deleter = allocator::free);
+      Deleter deleter = allocator::free);
 
   /** Assignment to rvalue does not compile. */
   array& operator=(const array& other) && = delete;
@@ -96,7 +99,7 @@ class array {
   }
 
   /** The shape of the array as a vector of integers. */
-  const std::vector<int>& shape() const {
+  const Shape& shape() const {
     return array_desc_->shape;
   }
 
@@ -105,12 +108,12 @@ class array {
    *
    *  This function supports negative indexing and provides
    *  bounds checking. */
-  int shape(int dim) const {
+  auto shape(int dim) const {
     return shape().at(dim < 0 ? dim + ndim() : dim);
   }
 
   /** The strides of the array. */
-  const std::vector<size_t>& strides() const {
+  const Strides& strides() const {
     return array_desc_->strides;
   }
 
@@ -119,7 +122,7 @@ class array {
    *
    *  This function supports negative indexing and provides
    *  bounds checking. */
-  size_t strides(int dim) const {
+  auto strides(int dim) const {
     return strides().at(dim < 0 ? dim + ndim() : dim);
   }
 
@@ -184,13 +187,13 @@ class array {
    */
 
   array(
-      std::vector<int> shape,
+      Shape shape,
       Dtype dtype,
       std::shared_ptr<Primitive> primitive,
       std::vector<array> inputs);
 
   static std::vector<array> make_arrays(
-      std::vector<std::vector<int>> shapes,
+      std::vector<Shape> shapes,
       const std::vector<Dtype>& dtypes,
       const std::shared_ptr<Primitive>& primitive,
       const std::vector<array>& inputs);
@@ -207,8 +210,8 @@ class array {
 
   struct Data {
     allocator::Buffer buffer;
-    deleter_t d;
+    Deleter d;
-    Data(allocator::Buffer buffer, deleter_t d = allocator::free)
+    Data(allocator::Buffer buffer, Deleter d = allocator::free)
         : buffer(buffer), d(d) {}
     // Not copyable
     Data(const Data& d) = delete;
@@ -397,18 +400,18 @@ class array {
   // Check if the array is a tracer array
   bool is_tracer() const;
 
-  void set_data(allocator::Buffer buffer, deleter_t d = allocator::free);
+  void set_data(allocator::Buffer buffer, Deleter d = allocator::free);
 
   void set_data(
       allocator::Buffer buffer,
       size_t data_size,
-      std::vector<size_t> strides,
+      Strides strides,
       Flags flags,
-      deleter_t d = allocator::free);
+      Deleter d = allocator::free);
 
   void copy_shared_buffer(
       const array& other,
-      const std::vector<size_t>& strides,
+      const Strides& strides,
       Flags flags,
       size_t data_size,
       size_t offset = 0);
@@ -417,7 +420,7 @@ class array {
 
   void move_shared_buffer(
       array other,
-      const std::vector<size_t>& strides,
+      const Strides& strides,
       Flags flags,
       size_t data_size,
       size_t offset = 0);
@@ -436,8 +439,8 @@ class array {
   void init(const It src);
 
   struct ArrayDesc {
-    std::vector<int> shape;
+    Shape shape;
-    std::vector<size_t> strides;
+    Strides strides;
     size_t size;
     Dtype dtype;
     std::shared_ptr<Primitive> primitive;
@@ -471,10 +474,10 @@ class array {
     // The arrays position in the output list
     uint32_t position{0};
 
-    explicit ArrayDesc(std::vector<int> shape, Dtype dtype);
+    explicit ArrayDesc(Shape shape, Dtype dtype);
 
     explicit ArrayDesc(
-        std::vector<int> shape,
+        Shape shape,
         Dtype dtype,
         std::shared_ptr<Primitive> primitive,
         std::vector<array> inputs);
@@ -502,7 +505,7 @@ array::array(T val, Dtype dtype /* = TypeToDtype<T>() */)
 template <typename It>
 array::array(
   It data,
-  std::vector<int> shape,
+  Shape shape,
   Dtype dtype /* = TypeToDtype<typename std::iterator_traits<It>::value_type>() */) :
     array_desc_(std::make_shared<ArrayDesc>(std::move(shape), dtype)) {
   init(data);
@@ -521,7 +524,7 @@ array::array(
 template <typename T>
 array::array(
     std::initializer_list<T> data,
-    std::vector<int> shape,
+    Shape shape,
     Dtype dtype /* = TypeToDtype<T>() */)
     : array_desc_(std::make_shared<ArrayDesc>(std::move(shape), dtype)) {
   if (data.size() != size()) {

mlx/backend/common/arg_reduce.cpp

@@ -13,8 +13,8 @@ template <typename InT, typename OpT>
 void arg_reduce(const array& in, array& out, const OpT& op, int axis) {
   auto axis_size = in.shape()[axis];
   auto axis_stride = in.strides()[axis];
-  std::vector<size_t> strides = in.strides();
+  Strides strides = in.strides();
-  std::vector<int> shape = in.shape();
+  Shape shape = in.shape();
   strides.erase(strides.begin() + axis);
   shape.erase(shape.begin() + axis);
   for (uint32_t i = 0; i < out.size(); ++i) {

mlx/backend/common/binary.h

@@ -178,10 +178,10 @@ void binary_op_dims(
     const T* b,
     U* out,
     Op op,
-    const std::vector<int>& shape,
+    const Shape& shape,
-    const std::vector<size_t>& a_strides,
+    const Strides& a_strides,
-    const std::vector<size_t>& b_strides,
+    const Strides& b_strides,
-    const std::vector<size_t>& out_strides,
+    const Strides& out_strides,
     int axis) {
   auto stride_a = a_strides[axis];
   auto stride_b = b_strides[axis];
@@ -212,10 +212,10 @@ void binary_op_dispatch_dims(
     array& out,
     Op op,
     int dim,
-    const std::vector<int>& shape,
+    const Shape& shape,
-    const std::vector<size_t>& a_strides,
+    const Strides& a_strides,
-    const std::vector<size_t>& b_strides,
+    const Strides& b_strides,
-    const std::vector<size_t>& out_strides) {
+    const Strides& out_strides) {
   const T* a_ptr = a.data<T>();
   const T* b_ptr = b.data<T>();
   U* out_ptr = out.data<U>();
@@ -258,10 +258,10 @@ void binary_op_dispatch_dims(
       return;
   }
 
-  ContiguousIterator<size_t> a_it(shape, a_strides, dim - 3);
+  ContiguousIterator a_it(shape, a_strides, dim - 3);
-  ContiguousIterator<size_t> b_it(shape, b_strides, dim - 3);
+  ContiguousIterator b_it(shape, b_strides, dim - 3);
-  size_t stride = out_strides[dim - 4];
+  auto stride = out_strides[dim - 4];
-  for (size_t elem = 0; elem < a.size(); elem += stride) {
+  for (int64_t elem = 0; elem < a.size(); elem += stride) {
     binary_op_dims<T, U, Op, 3, Strided>(
         a_ptr + a_it.loc,
         b_ptr + b_it.loc,
@@ -327,7 +327,7 @@ void binary_op(
   const auto& strides = new_strides[2];
 
   // Get the left-most dim such that the array is row contiguous after
-  auto leftmost_rc_dim = [&strides](const std::vector<size_t>& arr_strides) {
+  auto leftmost_rc_dim = [&strides](const auto& arr_strides) {
     int d = arr_strides.size() - 1;
     for (; d >= 0 && arr_strides[d] == strides[d]; d--) {
     }
@@ -337,7 +337,7 @@ void binary_op(
   auto b_rc_dim = leftmost_rc_dim(b_strides);
 
   // Get the left-most dim such that the array is a broadcasted "scalar" after
-  auto leftmost_s_dim = [](const std::vector<size_t>& arr_strides) {
+  auto leftmost_s_dim = [](const auto& arr_strides) {
     int d = arr_strides.size() - 1;
     for (; d >= 0 && arr_strides[d] == 0; d--) {
     }

mlx/backend/common/binary_two.h

@@ -16,10 +16,10 @@ void binary_op_dims(
     U* out_a,
     U* out_b,
     Op op,
-    const std::vector<int>& shape,
+    const Shape& shape,
-    const std::vector<size_t>& a_strides,
+    const Strides& a_strides,
-    const std::vector<size_t>& b_strides,
+    const Strides& b_strides,
-    const std::vector<size_t>& out_strides,
+    const Strides& out_strides,
     int axis) {
   auto stride_a = a_strides[axis];
   auto stride_b = b_strides[axis];
@@ -96,9 +96,9 @@ void binary_op_dispatch_dims(
       return;
   }
 
-  ContiguousIterator<size_t> a_it(shape, a_strides, ndim - 2);
+  ContiguousIterator a_it(shape, a_strides, ndim - 2);
-  ContiguousIterator<size_t> b_it(shape, b_strides, ndim - 2);
+  ContiguousIterator b_it(shape, b_strides, ndim - 2);
-  size_t stride = out_strides[ndim - 3];
+  auto stride = out_strides[ndim - 3];
   for (size_t elem = 0; elem < a.size(); elem += stride) {
     binary_op_dims<T, U, Op, 2>(
         a_ptr + a_it.loc,

mlx/backend/common/common.cpp

@@ -39,7 +39,7 @@ void AsStrided::eval(const std::vector<array>& inputs, array& out) {
   // rely on data_size anyway.
   size_t data_size = out.size();
 
-  return out.copy_shared_buffer(in, strides_, flags, data_size, offset_);
+  return move_or_copy(in, out, strides_, flags, data_size, offset_);
 }
 
 void Broadcast::eval(const std::vector<array>& inputs, array& out) {
@@ -49,7 +49,7 @@ void Broadcast::eval(const std::vector<array>& inputs, array& out) {
     out.set_data(nullptr);
     return;
   }
-  std::vector<size_t> strides(out.ndim(), 0);
+  Strides strides(out.ndim(), 0);
   int diff = out.ndim() - in.ndim();
   for (int i = in.ndim() - 1; i >= 0; --i) {
     strides[i + diff] = (in.shape()[i] == 1) ? 0 : in.strides()[i];
@@ -58,12 +58,12 @@ void Broadcast::eval(const std::vector<array>& inputs, array& out) {
   if (out.size() > in.size()) {
     flags.row_contiguous = flags.col_contiguous = false;
   }
-  out.copy_shared_buffer(in, strides, flags, in.data_size());
+  move_or_copy(in, out, strides, flags, in.data_size());
 }
 
 void Copy::eval(const std::vector<array>& inputs, array& out) {
   assert(inputs.size() == 1);
-  out.copy_shared_buffer(inputs[0]);
+  move_or_copy(inputs[0], out);
 }
 
 void CustomTransforms::eval(
@@ -72,7 +72,7 @@ void CustomTransforms::eval(
   assert(inputs.size() > outputs.size());
   for (int i = 0, j = inputs.size() - outputs.size(); i < outputs.size();
        i++, j++) {
-    outputs[i].copy_shared_buffer(inputs[j]);
+    move_or_copy(inputs[j], outputs[i]);
   }
 }
 
@@ -81,7 +81,7 @@ void Depends::eval(
     std::vector<array>& outputs) {
   assert(inputs.size() > outputs.size());
   for (int i = 0; i < outputs.size(); i++) {
-    outputs[i].copy_shared_buffer(inputs[i]);
+    move_or_copy(inputs[i], outputs[i]);
   }
 }
 
@@ -141,7 +141,7 @@ void NumberOfElements::eval(const std::vector<array>& inputs, array& out) {
   }
 }
 
-std::pair<bool, std::vector<size_t>> Reshape::prepare_reshape(
+std::pair<bool, Strides> Reshape::prepare_reshape(
     const array& in,
     const array& out) {
   // Special case for empty arrays or row contiguous arrays
@@ -151,8 +151,7 @@ std::pair<bool, std::vector<size_t>> Reshape::prepare_reshape(
 
   // Special case for scalars
   if (in.ndim() == 0) {
-    std::vector<size_t> out_strides(out.ndim(), 0);
+    return {false, Strides(out.ndim(), 0)};
-    return {false, out_strides};
   }
 
   // Firstly let's collapse all the contiguous dimensions of the input
@@ -160,7 +159,7 @@ std::pair<bool, std::vector<size_t>> Reshape::prepare_reshape(
 
   // If shapes fit exactly in the contiguous dims then no copy is necessary so
   // let's check.
-  std::vector<size_t> out_strides;
+  Strides out_strides;
   bool copy_necessary = false;
   int j = 0;
   for (int i = 0; i < out.ndim(); i++) {
@@ -183,7 +182,7 @@ std::pair<bool, std::vector<size_t>> Reshape::prepare_reshape(
 
 void Reshape::shared_buffer_reshape(
     const array& in,
-    const std::vector<size_t>& out_strides,
+    const Strides& out_strides,
     array& out) {
   auto flags = in.flags();
   if (flags.row_contiguous) {
@@ -194,7 +193,7 @@ void Reshape::shared_buffer_reshape(
     auto max_dim = std::max_element(out.shape().begin(), out.shape().end());
     flags.col_contiguous = out.size() <= 1 || out.size() == *max_dim;
   }
-  out.copy_shared_buffer(in, out_strides, flags, in.data_size());
+  move_or_copy(in, out, out_strides, flags, in.data_size());
 }
 
 void Split::eval(
@@ -249,26 +248,14 @@ void Split::eval(
   }
 }
 
-std::tuple<int64_t, std::vector<int64_t>> SliceUpdate::prepare_slice(
-    const array& in) {
-  int64_t data_offset = 0;
-  std::vector<int64_t> inp_strides(in.ndim(), 0);
-  for (int i = 0; i < in.ndim(); ++i) {
-    data_offset += start_indices_[i] * in.strides()[i];
-    inp_strides[i] = in.strides()[i] * strides_[i];
-  }
-
-  return std::make_tuple(data_offset, inp_strides);
-}
-
 void StopGradient::eval(const std::vector<array>& inputs, array& out) {
   assert(inputs.size() == 1);
-  out.copy_shared_buffer(inputs[0]);
+  move_or_copy(inputs[0], out);
 }
 
 void Transpose::eval(const std::vector<array>& inputs, array& out) {
   assert(inputs.size() == 1);
-  std::vector<size_t> out_strides(out.ndim());
+  Strides out_strides(out.ndim());
   auto& in = inputs[0];
   for (int ax = 0; ax < axes_.size(); ++ax) {
     out_strides[ax] = in.strides()[axes_[ax]];
@@ -285,8 +272,8 @@ void Transpose::eval(const std::vector<array>& inputs, array& out) {
   //   true, they stay true)
   auto flags = in.flags();
   if (flags.contiguous && in.data_size() == in.size()) {
-    size_t f_stride = 1;
+    int64_t f_stride = 1;
-    size_t b_stride = 1;
+    int64_t b_stride = 1;
     flags.col_contiguous = true;
     flags.row_contiguous = true;
     for (int i = 0, ri = out.ndim() - 1; i < out.ndim(); ++i, --ri) {
@@ -297,7 +284,7 @@ void Transpose::eval(const std::vector<array>& inputs, array& out) {
       b_stride *= out.shape(ri);
     }
   }
-  out.copy_shared_buffer(in, out_strides, flags, in.data_size());
+  move_or_copy(in, out, out_strides, flags, in.data_size());
 }
 
 } // namespace mlx::core

mlx/backend/common/compiled.cpp

@@ -165,7 +165,7 @@ void compiled_allocate_outputs(
     bool move_buffers /* = false */) {
   if (contiguous) {
     int o = 0;
-    std::vector<size_t> strides;
+    Strides strides;
     size_t data_size;
     array::Flags flags;
     for (int i = 0; i < inputs.size() && o < outputs.size(); ++i) {

mlx/backend/common/compiled_cpu.cpp

@@ -47,7 +47,7 @@ bool compile_available_for_device(const Device& device) {
 } // namespace detail
 
 std::string get_temp_file(const std::string& name) {
-  return std::filesystem::temp_directory_path().append(name);
+  return std::filesystem::temp_directory_path().append(name).string();
 }
 
 // Return a pointer to a compiled function
@@ -279,7 +279,7 @@ void Compiled::eval_cpu(
 
   // Figure out which kernel we are using
   auto& shape = outputs[0].shape();
-  bool contiguous = compiled_check_contiguity(inputs, shape);
+  auto contiguous = compiled_check_contiguity(inputs, shape);
 
   // Handle all broadcasting and collect function input arguments
   std::vector<void*> args;

mlx/backend/common/conv.cpp

@@ -746,9 +746,9 @@ void explicit_gemm_conv_1D_cpu(
   copy_inplace(in, in_padded_slice, CopyType::GeneralGeneral);
 
   // Make strided view
-  std::vector<int> strided_shape = {N, oH, wH, C};
+  Shape strided_shape = {N, oH, wH, C};
 
-  std::vector<size_t> strided_strides = {
+  Strides strided_strides = {
       in_padded.strides()[0],
       in_padded.strides()[1] * wt_strides[0],
       in_padded.strides()[1],
@@ -865,9 +865,9 @@ void explicit_gemm_conv_2D_cpu(
   copy_inplace(in, in_padded_slice, CopyType::GeneralGeneral);
 
   // Make strided view
-  std::vector<int> strided_shape = {N, oH, oW, wH, wW, C};
+  Shape strided_shape = {N, oH, oW, wH, wW, C};
 
-  std::vector<size_t> strided_strides = {
+  Strides strided_strides = {
       in_padded.strides()[0],
       in_padded.strides()[1] * wt_strides[0],
       in_padded.strides()[2] * wt_strides[1],
@@ -974,7 +974,7 @@ void explicit_gemm_conv_ND_cpu(
   copy_inplace(in, in_padded_slice, CopyType::GeneralGeneral);
 
   // Make strided view
-  std::vector<int> strided_shape(oDim.size() + wDim.size() + 2);
+  Shape strided_shape(oDim.size() + wDim.size() + 2);
   strided_shape.front() = N;
   for (size_t i = 0; i < oDim.size(); i++) {
     strided_shape[i + 1] = oDim[i];
@@ -984,7 +984,7 @@ void explicit_gemm_conv_ND_cpu(
   }
   strided_shape.back() = C;
 
-  std::vector<size_t> strided_strides(in.shape().size() * 2 - 2);
+  Strides strided_strides(in.shape().size() * 2 - 2);
   strided_strides[0] = in_padded.strides()[0];
   for (size_t i = 0; i < wt_strides.size(); i++) {
     strided_strides[i + 1] = in_padded.strides()[i + 1] * wt_strides[i];
@@ -1000,7 +1000,7 @@ void explicit_gemm_conv_ND_cpu(
       in_padded, strided_strides, flags, in_strided_view.size(), 0);
 
   // Materialize strided view
-  std::vector<int> strided_reshape = {N, C};
+  Shape strided_reshape = {N, C};
   for (const auto& o : oDim) {
     strided_reshape[0] *= o;
   }

mlx/backend/common/copy.cpp

@@ -26,13 +26,13 @@ void copy_vector(const array& src, array& dst) {
   std::copy(src_ptr, src_ptr + src.data_size(), dst_ptr);
 }
 
-template <typename SrcT, typename DstT, typename StrideT, int D>
+template <typename SrcT, typename DstT, int D>
 inline void copy_dims(
     const SrcT* src,
     DstT* dst,
-    const std::vector<int>& shape,
+    const Shape& shape,
-    const std::vector<StrideT>& i_strides,
+    const Strides& i_strides,
-    const std::vector<StrideT>& o_strides,
+    const Strides& o_strides,
     int axis) {
   auto stride_src = i_strides[axis];
   auto stride_dst = o_strides[axis];
@@ -40,7 +40,7 @@ inline void copy_dims(
 
   for (int i = 0; i < N; i++) {
     if constexpr (D > 1) {
-      copy_dims<SrcT, DstT, StrideT, D - 1>(
+      copy_dims<SrcT, DstT, D - 1>(
           src, dst, shape, i_strides, o_strides, axis + 1);
     } else {
       *dst = static_cast<DstT>(*src);
@@ -50,13 +50,13 @@ inline void copy_dims(
   }
 }
 
-template <typename SrcT, typename DstT, typename StrideT>
+template <typename SrcT, typename DstT>
 void copy_general_general(
     const array& src,
     array& dst,
-    const std::vector<int>& data_shape,
+    const Shape& data_shape,
-    const std::vector<StrideT>& i_strides,
+    const Strides& i_strides,
-    const std::vector<StrideT>& o_strides,
+    const Strides& o_strides,
     int64_t i_offset,
     int64_t o_offset) {
   if (data_shape.empty()) {
@@ -65,30 +65,30 @@ void copy_general_general(
     *dst_ptr = val;
     return;
   }
-  auto [shape, strides] = collapse_contiguous_dims(
+  auto [shape, strides] =
-      data_shape, std::vector<std::vector<StrideT>>{i_strides, o_strides});
+      collapse_contiguous_dims(data_shape, {i_strides, o_strides});
   auto src_ptr = src.data<SrcT>() + i_offset;
   auto dst_ptr = dst.data<DstT>() + o_offset;
   int ndim = shape.size();
   if (ndim == 1) {
-    copy_dims<SrcT, DstT, StrideT, 1>(
+    copy_dims<SrcT, DstT, 1>(
         src_ptr, dst_ptr, shape, strides[0], strides[1], 0);
     return;
   } else if (ndim == 2) {
-    copy_dims<SrcT, DstT, StrideT, 2>(
+    copy_dims<SrcT, DstT, 2>(
         src_ptr, dst_ptr, shape, strides[0], strides[1], 0);
     return;
   } else if (ndim == 3) {
-    copy_dims<SrcT, DstT, StrideT, 3>(
+    copy_dims<SrcT, DstT, 3>(
         src_ptr, dst_ptr, shape, strides[0], strides[1], 0);
     return;
   }
-  ContiguousIterator<StrideT> in(shape, strides[0], ndim - 3);
+  ContiguousIterator in(shape, strides[0], ndim - 3);
-  ContiguousIterator<StrideT> out(shape, strides[1], ndim - 3);
+  ContiguousIterator out(shape, strides[1], ndim - 3);
-  StrideT stride = std::accumulate(
+  auto stride = std::accumulate(
-      shape.end() - 3, shape.end(), 1, std::multiplies<StrideT>());
+      shape.end() - 3, shape.end(), 1, std::multiplies<int64_t>());
-  for (StrideT elem = 0; elem < src.size(); elem += stride) {
+  for (int64_t elem = 0; elem < src.size(); elem += stride) {
-    copy_dims<SrcT, DstT, StrideT, 3>(
+    copy_dims<SrcT, DstT, 3>(
         src_ptr + in.loc,
         dst_ptr + out.loc,
         shape,
@@ -102,37 +102,37 @@ void copy_general_general(
 
 template <typename SrcT, typename DstT>
 inline void copy_general_general(const array& src, array& dst) {
-  copy_general_general<SrcT, DstT, size_t>(
+  copy_general_general<SrcT, DstT>(
       src, dst, src.shape(), src.strides(), dst.strides(), 0, 0);
 }
 
-template <typename SrcT, typename DstT, typename StrideT>
+template <typename SrcT, typename DstT>
 void copy_general(
     const array& src,
     array& dst,
-    const std::vector<int>& data_shape,
+    const Shape& data_shape,
-    const std::vector<StrideT>& i_strides,
+    const Strides& i_strides,
-    const std::vector<StrideT>&,
+    const Strides&,
     int64_t i_offset,
     int64_t o_offset) {
-  copy_general_general<SrcT, DstT, StrideT>(
+  copy_general_general<SrcT, DstT>(
       src,
       dst,
       data_shape,
       i_strides,
-      make_contiguous_strides<StrideT>(data_shape),
+      make_contiguous_strides(data_shape),
       i_offset,
       o_offset);
 }
 
 template <typename SrcT, typename DstT>
 inline void copy_general(const array& src, array& dst) {
-  copy_general_general<SrcT, DstT, size_t>(
+  copy_general_general<SrcT, DstT>(
       src,
       dst,
       src.shape(),
       src.strides(),
-      make_contiguous_strides<size_t>(src.shape()),
+      make_contiguous_strides(src.shape()),
       0,
       0);
 }
@@ -282,13 +282,12 @@ void copy(const array& src, array& dst, CopyType ctype) {
   copy_inplace(src, dst, ctype);
 }
 
-template <typename StrideT>
 void copy_inplace(
     const array& src,
     array& dst,
-    const std::vector<int>& data_shape,
+    const Shape& data_shape,
-    const std::vector<StrideT>& i_strides,
+    const Strides& i_strides,
-    const std::vector<StrideT>& o_strides,
+    const Strides& o_strides,
     int64_t i_offset,
     int64_t o_offset,
     CopyType ctype) {
@@ -311,24 +310,4 @@ void copy_inplace(
   }
 }
 
-template void copy_inplace<size_t>(
-    const array& src,
-    array& dst,
-    const std::vector<int>& data_shape,
-    const std::vector<size_t>& i_strides,
-    const std::vector<size_t>& o_strides,
-    int64_t i_offset,
-    int64_t o_offset,
-    CopyType ctype);
-
-template void copy_inplace<int64_t>(
-    const array& src,
-    array& dst,
-    const std::vector<int>& data_shape,
-    const std::vector<int64_t>& i_strides,
-    const std::vector<int64_t>& o_strides,
-    int64_t i_offset,
-    int64_t o_offset,
-    CopyType ctype);
-
 } // namespace mlx::core

mlx/backend/common/copy.h

@@ -26,13 +26,12 @@ enum class CopyType {
 void copy(const array& src, array& dst, CopyType ctype);
 void copy_inplace(const array& src, array& dst, CopyType ctype);
 
-template <typename stride_t>
 void copy_inplace(
     const array& src,
     array& dst,
-    const std::vector<int>& data_shape,
+    const Shape& data_shape,
-    const std::vector<stride_t>& i_strides,
+    const Strides& i_strides,
-    const std::vector<stride_t>& o_strides,
+    const Strides& o_strides,
     int64_t i_offset,
     int64_t o_offset,
     CopyType ctype);

mlx/backend/common/default_primitives.cpp

@@ -130,7 +130,7 @@ inline void matmul_common_general(
     } else {
       array arr_copy(arr.shape(), arr.dtype(), nullptr, {});
       copy(arr, arr_copy, CopyType::General);
-      size_t stx = arr.shape(-1);
+      stx = arr.shape(-1);
       return std::make_tuple(false, stx, arr_copy);
     }
   };

mlx/backend/common/indexing.cpp

@@ -32,7 +32,7 @@ void gather(
     const std::vector<array>& inds,
     array& out,
     const std::vector<int>& axes,
-    const std::vector<int>& slice_sizes) {
+    const Shape& slice_sizes) {
   // If the array is row contiguous then we can do a contiguous copy given
   // two conditions on the slice size:
   // - Any number of leading ones in the slice sizes are allowed
@@ -80,11 +80,10 @@ void gather(
   T* dst_ptr = out.data<T>();
   size_t out_idx = 0;
 
-  std::vector<ContiguousIterator<size_t>> its(inds.begin(), inds.end());
+  std::vector<ContiguousIterator> its(inds.begin(), inds.end());
-  ContiguousIterator<size_t> src_it;
+  ContiguousIterator src_it;
   if (!can_copy && src.ndim() > 0) {
-    src_it = std::move(
+    src_it = ContiguousIterator(slice_sizes, src.strides(), src.ndim());
-        ContiguousIterator<size_t>(slice_sizes, src.strides(), src.ndim()));
   }
   for (int idx = 0; idx < ind_size; idx++) {
     size_t src_idx = 0;
@@ -119,7 +118,7 @@ void dispatch_gather(
     const std::vector<array>& inds,
     array& out,
     const std::vector<int>& axes,
-    const std::vector<int>& size) {
+    const Shape& size) {
   switch (out.dtype()) {
     case bool_:
       gather<bool, IdxT>(src, inds, out, axes, size);
@@ -223,16 +222,16 @@ void scatter(
   auto inds_ndim = updates.ndim() - out.ndim();
   size_t n_updates = nind ? inds[0].size() : 1;
 
-  std::vector<int> update_shape(
+  Shape update_shape(
       updates.shape().begin() + inds_ndim, updates.shape().end());
   size_t update_size = 1;
   for (auto us : update_shape) {
     update_size *= us;
   }
 
-  std::vector<ContiguousIterator<size_t>> its(inds.begin(), inds.end());
+  std::vector<ContiguousIterator> its(inds.begin(), inds.end());
-  ContiguousIterator<size_t> update_it(updates);
+  ContiguousIterator update_it(updates);
-  ContiguousIterator<size_t> out_it(update_shape, out.strides(), out.ndim());
+  ContiguousIterator out_it(update_shape, out.strides(), out.ndim());
 
   for (int i = 0; i < n_updates; ++i) {
     size_t out_offset = 0;

mlx/backend/common/lapack.h

@@ -2,6 +2,15 @@
 
 #pragma once
 
+// Required for Visual Studio.
+// https://github.com/OpenMathLib/OpenBLAS/blob/develop/docs/install.md
+#ifdef _MSC_VER
+#include <complex>
+#define LAPACK_COMPLEX_CUSTOM
+#define lapack_complex_float std::complex<float>
+#define lapack_complex_double std::complex<double>
+#endif
+
 #ifdef ACCELERATE_NEW_LAPACK
 #include <Accelerate/Accelerate.h>
 #else

mlx/backend/common/masked_mm.cpp

@@ -19,10 +19,10 @@ inline void mask_matrix(
     int block_size,
     const int X,
     const int Y,
-    const size_t X_data_str,
+    const int64_t X_data_str,
-    const size_t Y_data_str,
+    const int64_t Y_data_str,
-    const size_t X_mask_str,
+    const int64_t X_mask_str,
-    const size_t Y_mask_str,
+    const int64_t Y_mask_str,
     const size_t mask_offset) {
   int tX = (X + block_size - 1) / block_size;
   int tY = (Y + block_size - 1) / block_size;
@@ -84,7 +84,7 @@ void BlockMaskedMM::eval(const std::vector<array>& inputs, array& out) {
         } else {
           array arr_copy(arr.shape(), arr.dtype(), nullptr, {});
           copy(arr, arr_copy, CopyType::General);
-          size_t stx = arr.shape(-1);
+          int64_t stx = arr.shape(-1);
           return std::make_tuple(false, stx, arr_copy);
         }
       };
@@ -117,13 +117,13 @@ void BlockMaskedMM::eval(const std::vector<array>& inputs, array& out) {
                        int Y,
                        size_t X_data_str,
                        size_t Y_data_str) {
-    size_t mask_offset = elem_to_loc(
+    auto mask_offset = elem_to_loc(
         mask.shape(-1) * mask.shape(-2) * batch_idx,
         mask.shape(),
         mask.strides());
 
-    size_t X_mask_str = mask.strides()[mask.ndim() - 2];
+    auto X_mask_str = mask.strides()[mask.ndim() - 2];
-    size_t Y_mask_str = mask.strides()[mask.ndim() - 1];
+    auto Y_mask_str = mask.strides()[mask.ndim() - 1];
 
     if (mask.dtype() == bool_) {
       return mask_matrix(
@@ -230,7 +230,7 @@ void GatherMM::eval(const std::vector<array>& inputs, array& out) {
     } else {
       array arr_copy(arr.shape(), arr.dtype(), nullptr, {});
       copy(arr, arr_copy, CopyType::General);
-      size_t stx = arr.shape(-1);
+      int64_t stx = arr.shape(-1);
       return std::make_tuple(false, stx, arr_copy);
     }
   };
@@ -262,13 +262,13 @@ void GatherMM::eval(const std::vector<array>& inputs, array& out) {
   auto& lhs_indices = inputs[2];
   auto& rhs_indices = inputs[3];
 
-  std::vector<int> batch_shape = get_batch_dims(out.shape());
+  auto batch_shape = get_batch_dims(out.shape());
   int batch_ndim = batch_shape.size();
 
-  std::vector<int> batch_shape_A = get_batch_dims(a.shape());
+  auto batch_shape_A = get_batch_dims(a.shape());
-  std::vector<size_t> batch_strides_A = get_batch_dims(a.strides());
+  auto batch_strides_A = get_batch_dims(a.strides());
-  std::vector<int> batch_shape_B = get_batch_dims(b.shape());
+  auto batch_shape_B = get_batch_dims(b.shape());
-  std::vector<size_t> batch_strides_B = get_batch_dims(b.strides());
+  auto batch_strides_B = get_batch_dims(b.strides());
 
   const uint32_t* lhs_indices_ptr = lhs_indices.data<uint32_t>();
   const uint32_t* rhs_indices_ptr = rhs_indices.data<uint32_t>();

mlx/backend/common/ops.h

@@ -500,7 +500,12 @@ struct Equal {
 struct NaNEqual {
   template <typename T>
   bool operator()(T x, T y) {
-    return x == y || (std::isnan(x) && std::isnan(y));
+    if constexpr (std::is_integral_v<T>) {
+      // isnan always returns false for integers, and MSVC refuses to compile.
+      return x == y;
+    } else {
+      return x == y || (std::isnan(x) && std::isnan(y));
+    }
   }
 };
 

mlx/backend/common/primitives.cpp

@@ -159,6 +159,17 @@ void Conjugate::eval(const std::vector<array>& inputs, array& out) {
   }
 }
 
+void Contiguous::eval_cpu(const std::vector<array>& inputs, array& out) {
+  assert(inputs.size() == 1);
+  auto& in = inputs[0];
+  if (in.flags().row_contiguous ||
+      (allow_col_major_ && in.flags().col_contiguous)) {
+    out.copy_shared_buffer(in);
+  } else {
+    copy(in, out, CopyType::General);
+  }
+}
+
 void Cos::eval(const std::vector<array>& inputs, array& out) {
   assert(inputs.size() == 1);
   const auto& in = inputs[0];
@@ -487,14 +498,15 @@ void Slice::eval(const std::vector<array>& inputs, array& out) {
   auto& in = inputs[0];
 
   // Calculate out strides, initial offset and if copy needs to be made
-  auto [copy_needed, data_offset, inp_strides] =
+  auto [data_offset, inp_strides] = prepare_slice(in, start_indices_, strides_);
-      prepare_slice(in, start_indices_, strides_);
+  auto copy_needed = std::any_of(
+      strides_.begin(), strides_.end(), [](auto i) { return i < 0; });
 
   // Do copy if needed
   if (copy_needed) {
     out.set_data(allocator::malloc_or_wait(out.nbytes()));
-    std::vector<int64_t> ostrides{out.strides().begin(), out.strides().end()};
+    Strides ostrides{out.strides().begin(), out.strides().end()};
-    copy_inplace<int64_t>(
+    copy_inplace(
         /* const array& src = */ in,
         /* array& dst = */ out,
         /* const std::vector<int>& data_shape = */ out.shape(),
@@ -512,7 +524,7 @@ void Slice::eval(const std::vector<array>& inputs, array& out) {
       }
     }
     size_t data_size = data_end - data_offset;
-    std::vector<size_t> ostrides{inp_strides.begin(), inp_strides.end()};
+    Strides ostrides{inp_strides.begin(), inp_strides.end()};
     shared_buffer_slice(in, ostrides, data_offset, data_size, out);
   }
 }
@@ -539,11 +551,11 @@ void SliceUpdate::eval(const std::vector<array>& inputs, array& out) {
   copy(in, out, in.data_size() == 1 ? CopyType::Scalar : ctype);
 
   // Calculate out strides, initial offset and if copy needs to be made
-  auto [data_offset, out_strides] = prepare_slice(out);
+  auto [data_offset, out_strides] = prepare_slice(in, start_indices_, strides_);
 
   // Do copy
-  std::vector<int64_t> upd_strides{upd.strides().begin(), upd.strides().end()};
+  Strides upd_strides{upd.strides().begin(), upd.strides().end()};
-  copy_inplace<int64_t>(
+  copy_inplace(
       /* const array& src = */ upd,
       /* array& dst = */ out,
       /* const std::vector<int>& data_shape = */ upd.shape(),
@@ -606,7 +618,7 @@ void View::eval_cpu(const std::vector<array>& inputs, array& out) {
   if (ibytes == obytes || obytes < ibytes && in.strides().back() == 1 ||
       in.flags().row_contiguous) {
     auto strides = in.strides();
-    for (int i = 0; i < strides.size() - 1; ++i) {
+    for (int i = 0; i < static_cast<int>(strides.size()) - 1; ++i) {
       strides[i] *= ibytes;
       strides[i] /= obytes;
     }

mlx/backend/common/qrf.cpp

@@ -54,7 +54,7 @@ void qrf_impl(const array& a, array& q, array& r) {
   // Copy the input to be column contiguous
   flags.col_contiguous = num_matrices == 1;
   flags.row_contiguous = false;
-  std::vector<size_t> strides = in.strides();
+  auto strides = in.strides();
   strides[in.ndim() - 2] = 1;
   strides[in.ndim() - 1] = M;
   in.set_data(

mlx/backend/common/quantized.cpp

@@ -2,13 +2,38 @@
 
 #include <cassert>
 
-#include "mlx/backend/metal/copy.h"
+#include "mlx/backend/common/copy.h"
+#include "mlx/backend/common/ops.h"
+#include "mlx/fast_primitives.h"
 #include "mlx/primitives.h"
+#include "mlx/utils.h"
 
 namespace mlx::core {
 
 namespace {
 
+template <typename T, int bits>
+void extract_bits(const uint8_t* w_in, T* w_out) {
+  assert(bits == 3 || bits == 6);
+  if (bits == 3) {
+    w_out[0] = static_cast<T>(w_in[0] & 0x7);
+    w_out[1] = static_cast<T>((w_in[0] & 0x38) >> 3);
+    w_out[2] = static_cast<T>(((w_in[0] & 0xc0) >> 6) + ((w_in[1] & 0x1) << 2));
+    w_out[3] = static_cast<T>((w_in[1] & 0xe) >> 1);
+    w_out[4] = static_cast<T>((w_in[1] & 0x70) >> 4);
+    w_out[5] = static_cast<T>(((w_in[1] & 0x80) >> 7) + ((w_in[2] & 0x3) << 1));
+    w_out[6] = static_cast<T>((w_in[2] & 0x1c) >> 2);
+    w_out[7] = static_cast<T>((w_in[2] & 0xe0) >> 5);
+  } else if (bits == 6) {
+    w_out[0] = static_cast<T>(w_in[0] & 0x3f);
+    w_out[1] =
+        static_cast<T>(((w_in[0] >> 6) & 0x03) + ((w_in[1] & 0x0f) << 2));
+    w_out[2] =
+        static_cast<T>(((w_in[1] >> 4) & 0x0f) + ((w_in[2] & 0x03) << 4));
+    w_out[3] = static_cast<T>((w_in[2] >> 2) & 0x3f);
+  }
+}
+
 template <typename T, int bits, int group_size>
 void _qmm(
     T* result,
@@ -20,13 +45,12 @@ void _qmm(
     int N,
     int K) {
   constexpr int bitmask = (1 << bits) - 1;
-  constexpr int pack_factor = 32 / bits;
+  constexpr int pack_factor = bits == 3 ? 8 : bits == 6 ? 4 : 8 / bits;
+  constexpr int bytes_per_pack = (bits == 3 || bits == 6) ? 3 : 1;
   constexpr int packs_in_group = group_size / pack_factor;
-  const int Ng = N / group_size;
-  const int Nw = N / pack_factor;
 
   for (int m = 0; m < M; m++) {
-    const uint32_t* w_local = w;
+    const uint8_t* w_local = (const uint8_t*)w;
     const T* scales_local = scales;
     const T* biases_local = biases;
 
@@ -40,13 +64,25 @@ void _qmm(
         T scale = *scales_local++;
         T bias = *biases_local++;
         for (int ng = 0; ng < packs_in_group; ng++) {
-          uint32_t wi = *w_local++;
+          if (bits == 3 || bits == 6) {
-
+            T wl[pack_factor];
+            extract_bits<T, bits>(w_local, wl);
 #pragma clang loop unroll(full)
-          for (int p = 0; p < pack_factor; p++) {
+            for (int p = 0; p < pack_factor; p++) {
-            (*result_local++) +=
+              (*result_local++) += xi * (scale * wl[p] + bias);
-                xi * (scale * static_cast<T>(wi & bitmask) + bias);
+            }
-            wi >>= bits;
+            w_local += bytes_per_pack;
+
+          } else {
+            uint8_t wi = *w_local++;
+#pragma clang loop unroll(full)
+            for (int p = 0; p < pack_factor; p++) {
+              (*result_local++) +=
+                  xi * (scale * static_cast<T>(wi & bitmask) + bias);
+              if (bits != 8) {
+                wi >>= bits;
+              }
+            }
           }
         }
       }
@@ -67,13 +103,12 @@ void _qmm_t(
     int N,
     int K) {
   constexpr int bitmask = (1 << bits) - 1;
-  constexpr int pack_factor = 32 / bits;
+  constexpr int pack_factor = bits == 3 ? 8 : bits == 6 ? 4 : 8 / bits;
+  constexpr int bytes_per_pack = (bits == 3 || bits == 6) ? 3 : 1;
   constexpr int packs_in_group = group_size / pack_factor;
-  const int Kg = K / group_size;
-  const int Kw = K / pack_factor;
 
   for (int m = 0; m < M; m++) {
-    const uint32_t* w_local = w;
+    const uint8_t* w_local = (const uint8_t*)w;
     const T* scales_local = scales;
     const T* biases_local = biases;
 
@@ -85,12 +120,26 @@ void _qmm_t(
         T bias = *biases_local++;
 
         for (int kw = 0; kw < packs_in_group; kw++) {
-          uint32_t wi = *w_local++;
+          if (bits == 3 || bits == 6) {
-
+            T wl[pack_factor];
+            extract_bits<T, bits>(w_local, wl);
 #pragma clang loop unroll(full)
-          for (int p = 0; p < pack_factor; p++) {
+            for (int p = 0; p < pack_factor; p++) {
-            sum += (*x_local++) * (scale * static_cast<T>(wi & bitmask) + bias);
+              sum += x_local[p] * (scale * wl[p] + bias);
-            wi >>= bits;
+            }
+            w_local += bytes_per_pack;
+            x_local += pack_factor;
+
+          } else {
+            uint8_t wi = *w_local++;
+#pragma clang loop unroll(full)
+            for (int p = 0; p < pack_factor; p++) {
+              sum +=
+                  (*x_local++) * (scale * static_cast<T>(wi & bitmask) + bias);
+              if (bits != 8) {
+                wi >>= bits;
+              }
+            }
           }
         }
       }
@@ -102,6 +151,55 @@ void _qmm_t(
   }
 }
 
+template <typename T, int bits, int group_size>
+void _qmm_dispatch_transpose(
+    T* result,
+    const T* x,
+    const uint32_t* w,
+    const T* scales,
+    const T* biases,
+    int M,
+    int N,
+    int K,
+    bool transposed_w) {
+  if (transposed_w) {
+    return _qmm_t<T, bits, group_size>(result, x, w, scales, biases, M, N, K);
+  } else {
+    return _qmm<T, bits, group_size>(result, x, w, scales, biases, M, N, K);
+  }
+}
+
+template <typename T, int bits>
+void _qmm_dispatch_group(
+    T* result,
+    const T* x,
+    const uint32_t* w,
+    const T* scales,
+    const T* biases,
+    int M,
+    int N,
+    int K,
+    int group_size,
+    bool transposed_w) {
+  switch (group_size) {
+    case 32:
+      _qmm_dispatch_transpose<T, bits, 32>(
+          result, x, w, scales, biases, M, N, K, transposed_w);
+      break;
+    case 64:
+      _qmm_dispatch_transpose<T, bits, 64>(
+          result, x, w, scales, biases, M, N, K, transposed_w);
+      break;
+    case 128:
+      _qmm_dispatch_transpose<T, bits, 128>(
+          result, x, w, scales, biases, M, N, K, transposed_w);
+      break;
+    default:
+      throw std::invalid_argument(
+          "Quantization group size must be 32, 64 or 128.");
+  }
+}
+
 template <typename T>
 void _qmm_dispatch_typed(
     T* result,
@@ -116,79 +214,29 @@ void _qmm_dispatch_typed(
     int bits,
     bool transposed_w) {
   switch (bits) {
-    case 2: {
+    case 2:
-      switch (group_size) {
+      _qmm_dispatch_group<T, 2>(
-        case 32:
+          result, x, w, scales, biases, M, N, K, group_size, transposed_w);
-          if (transposed_w) {
+      break;
-            return _qmm_t<T, 2, 32>(result, x, w, scales, biases, M, N, K);
+    case 3:
-          } else {
+      _qmm_dispatch_group<T, 3>(
-            return _qmm<T, 2, 32>(result, x, w, scales, biases, M, N, K);
+          result, x, w, scales, biases, M, N, K, group_size, transposed_w);
-          }
+      break;
-        case 64:
+    case 4:
-          if (transposed_w) {
+      _qmm_dispatch_group<T, 4>(
-            return _qmm_t<T, 2, 64>(result, x, w, scales, biases, M, N, K);
+          result, x, w, scales, biases, M, N, K, group_size, transposed_w);
-          } else {
+      break;
-            return _qmm<T, 2, 64>(result, x, w, scales, biases, M, N, K);
+    case 6:
-          }
+      _qmm_dispatch_group<T, 6>(
-        case 128:
+          result, x, w, scales, biases, M, N, K, group_size, transposed_w);
-          if (transposed_w) {
+      break;
-            return _qmm_t<T, 2, 128>(result, x, w, scales, biases, M, N, K);
+    case 8:
-          } else {
+      _qmm_dispatch_group<T, 8>(
-            return _qmm<T, 2, 128>(result, x, w, scales, biases, M, N, K);
+          result, x, w, scales, biases, M, N, K, group_size, transposed_w);
-          }
+      break;
-      }
+    default:
-    }
+      throw std::invalid_argument("Quantization bits must be 2, 3, 4, 6 or 8.");
-    case 4: {
-      switch (group_size) {
-        case 32:
-          if (transposed_w) {
-            return _qmm_t<T, 4, 32>(result, x, w, scales, biases, M, N, K);
-          } else {
-            return _qmm<T, 4, 32>(result, x, w, scales, biases, M, N, K);
-          }
-        case 64:
-          if (transposed_w) {
-            return _qmm_t<T, 4, 64>(result, x, w, scales, biases, M, N, K);
-          } else {
-            return _qmm<T, 4, 64>(result, x, w, scales, biases, M, N, K);
-          }
-        case 128:
-          if (transposed_w) {
-            return _qmm_t<T, 4, 128>(result, x, w, scales, biases, M, N, K);
-          } else {
-            return _qmm<T, 4, 128>(result, x, w, scales, biases, M, N, K);
-          }
-      }
-    }
-    case 8: {
-      switch (group_size) {
-        case 32:
-          if (transposed_w) {
-            return _qmm_t<T, 8, 32>(result, x, w, scales, biases, M, N, K);
-          } else {
-            return _qmm<T, 8, 32>(result, x, w, scales, biases, M, N, K);
-          }
-        case 64:
-          if (transposed_w) {
-            return _qmm_t<T, 8, 64>(result, x, w, scales, biases, M, N, K);
-          } else {
-            return _qmm<T, 8, 64>(result, x, w, scales, biases, M, N, K);
-          }
-        case 128:
-          if (transposed_w) {
-            return _qmm_t<T, 8, 128>(result, x, w, scales, biases, M, N, K);
-          } else {
-            return _qmm<T, 8, 128>(result, x, w, scales, biases, M, N, K);
-          }
-      }
-    }
   }
-  std::ostringstream msg;
-  msg << "Quantization type not supported. Provided bits=" << bits
-      << " and group_size=" << group_size
-      << ". The supported options are bits in "
-      << "{2, 4, 8} and group_size in {64, 128}.";
-  throw std::invalid_argument(msg.str());
 }
 
 void _qmm_dispatch(
@@ -404,4 +452,114 @@ void GatherQMM::eval(const std::vector<array>& inputs, array& out) {
       transpose_);
 }
 
+template <typename T, typename U>
+void quantize(
+    const array& w_,
+    array& out_,
+    array& scales_,
+    array& biases_,
+    int bits,
+    int group_size) {
+  const T* w = w_.data<T>();
+
+  auto out = out_.data<U>();
+  T* scales = scales_.data<T>();
+  T* biases = biases_.data<T>();
+
+  T n_bins = (1 << bits) - 1;
+  T eps = 1e-7;
+  bool power_of_2_bits = is_power_of_2(bits);
+  int el_per_int = bits == 3 ? 8 : bits == 6 ? 4 : 32 / bits;
+  // For 3/6 bits we read 3 uint8s at a time instead of 1 uint32
+  int bytes_per_pack = power_of_2_bits ? 1 : 3;
+  int int_per_group = group_size * bytes_per_pack / el_per_int;
+  size_t n_groups = w_.size() / group_size;
+
+  for (size_t i = 0; i < n_groups; ++i) {
+    size_t w_idx = i * group_size;
+    T w_min = std::numeric_limits<float>::infinity();
+    T w_max = -w_min;
+    for (int j = 0; j < group_size; ++j) {
+      w_max = std::max(w_max, w[w_idx + j]);
+      w_min = std::min(w_min, w[w_idx + j]);
+    }
+    bool mask = std::abs(w_min) > std::abs(w_max);
+    T scale = std::max(T((w_max - w_min) / n_bins), eps);
+    scale = mask ? scale : -scale;
+
+    auto edge = mask ? w_min : w_max;
+    auto q0 = std::rint(edge / scale);
+    if (q0 == 0) {
+      scales[i] = scale;
+      biases[i] = 0;
+    } else {
+      scales[i] = edge / q0;
+      biases[i] = edge;
+    }
+    size_t out_idx = i * int_per_group;
+    for (int j = 0; j < int_per_group / bytes_per_pack; ++j) {
+      uint32_t out_el = 0;
+      for (int k = 0; k < el_per_int; ++k) {
+        T w_el = w[w_idx + j * el_per_int + k];
+        w_el = std::rint((w_el - biases[i]) / scales[i]);
+        w_el = std::min(std::max(w_el, T(0)), n_bins);
+        out_el |= static_cast<uint32_t>(w_el) << (k * bits);
+      }
+      if (power_of_2_bits) {
+        out[out_idx + j] = out_el;
+      } else {
+        out[out_idx + bytes_per_pack * j] = out_el & 0xff;
+        out[out_idx + bytes_per_pack * j + 1] = (out_el & 0xff00) >> 8;
+        out[out_idx + bytes_per_pack * j + 2] = (out_el & 0xff0000) >> 16;
+      }
+    }
+  }
+}
+
+void fast::AffineQuantize::eval_cpu(
+    const std::vector<array>& inputs,
+    std::vector<array>& outputs) {
+  auto ensure_row_contiguous = [](const array& arr) {
+    if (arr.flags().row_contiguous) {
+      return arr;
+    } else {
+      array arr_copy(arr.shape(), arr.dtype(), nullptr, {});
+      copy(arr, arr_copy, CopyType::General);
+      return arr_copy;
+    }
+  };
+  auto w = ensure_row_contiguous(inputs[0]);
+
+  auto& out = outputs[0];
+  out.set_data(allocator::malloc_or_wait(out.nbytes()));
+
+  auto& scales = outputs[1];
+  auto& biases = outputs[2];
+  scales.set_data(allocator::malloc_or_wait(scales.nbytes()));
+  biases.set_data(allocator::malloc_or_wait(biases.nbytes()));
+  if (w.dtype() == float16) {
+    if (is_power_of_2(bits_)) {
+      quantize<float16_t, uint32_t>(w, out, scales, biases, bits_, group_size_);
+    } else {
+      quantize<float16_t, uint8_t>(w, out, scales, biases, bits_, group_size_);
+    }
+  } else if (w.dtype() == bfloat16) {
+    if (is_power_of_2(bits_)) {
+      quantize<bfloat16_t, uint32_t>(
+          w, out, scales, biases, bits_, group_size_);
+    } else {
+      quantize<bfloat16_t, uint8_t>(w, out, scales, biases, bits_, group_size_);
+    }
+  } else if (w.dtype() == float32) {
+    if (is_power_of_2(bits_)) {
+      quantize<float, uint32_t>(w, out, scales, biases, bits_, group_size_);
+    } else {
+      quantize<float, uint8_t>(w, out, scales, biases, bits_, group_size_);
+    }
+  } else {
+    throw std::runtime_error(
+        "[fast::AffineQuantize::eval_cpu] Only supports floating point inputs");
+  }
+}
+
 } // namespace mlx::core

mlx/backend/common/reduce.cpp

@@ -120,65 +120,73 @@ struct MinReduce {
 };
 
 template <typename InT>
-void reduce_dispatch_out(
+void reduce_dispatch_and_or(
     const array& in,
     array& out,
     Reduce::ReduceType rtype,
     const std::vector<int>& axes) {
-  switch (rtype) {
+  if (rtype == Reduce::And) {
-    case Reduce::And: {
+    reduction_op<InT, bool>(in, out, axes, true, AndReduce());
-      reduction_op<InT, bool>(in, out, axes, true, AndReduce());
+  } else {
-      break;
+    reduction_op<InT, bool>(in, out, axes, false, OrReduce());
+  }
+}
+
+template <typename InT>
+void reduce_dispatch_sum_prod(
+    const array& in,
+    array& out,
+    Reduce::ReduceType rtype,
+    const std::vector<int>& axes) {
+  if (rtype == Reduce::Sum) {
+    auto op = [](auto y, auto x) { (*y) = (*y) + x; };
+    if constexpr (std::is_integral_v<InT> && sizeof(InT) <= 4) {
+      reduction_op<InT, int32_t>(in, out, axes, 0, op);
+    } else {
+      reduction_op<InT, InT>(in, out, axes, 0, op);
     }
-    case Reduce::Or: {
+  } else {
-      reduction_op<InT, bool>(in, out, axes, false, OrReduce());
+    auto op = [](auto y, auto x) { (*y) *= x; };
-      break;
+    if constexpr (std::is_integral_v<InT> && sizeof(InT) <= 4) {
-    }
+      reduction_op<InT, int32_t>(in, out, axes, 1, op);
-    case Reduce::Sum: {
+    } else {
-      auto op = [](auto y, auto x) { (*y) = (*y) + x; };
-      if (out.dtype() == int32) {
-        // special case since the input type can be bool
-        reduction_op<InT, int32_t>(in, out, axes, 0, op);
-      } else {
-        reduction_op<InT, InT>(in, out, axes, 0, op);
-      }
-      break;
-    }
-    case Reduce::Prod: {
-      auto op = [](auto y, auto x) { (*y) *= x; };
       reduction_op<InT, InT>(in, out, axes, 1, op);
-      break;
-    }
-    case Reduce::Max: {
-      auto init = Limits<InT>::min;
-      reduction_op<InT, InT>(in, out, axes, init, MaxReduce());
-      break;
-    }
-    case Reduce::Min: {
-      auto init = Limits<InT>::max;
-      reduction_op<InT, InT>(in, out, axes, init, MinReduce());
-      break;
     }
   }
 }
 
+template <typename InT>
+void reduce_dispatch_min_max(
+    const array& in,
+    array& out,
+    Reduce::ReduceType rtype,
+    const std::vector<int>& axes) {
+  if (rtype == Reduce::Max) {
+    auto init = Limits<InT>::min;
+    reduction_op<InT, InT>(in, out, axes, init, MaxReduce());
+  } else {
+    auto init = Limits<InT>::max;
+    reduction_op<InT, InT>(in, out, axes, init, MinReduce());
+  }
+}
+
 } // namespace
 
 void nd_loop(
     std::function<void(int)> callback,
-    const std::vector<int>& shape,
+    const Shape& shape,
-    const std::vector<size_t>& strides) {
+    const Strides& strides) {
   std::function<void(int, int)> loop_inner;
   loop_inner = [&](int dim, int offset) {
     if (dim < shape.size() - 1) {
-      int size = shape[dim];
+      auto size = shape[dim];
-      size_t stride = strides[dim];
+      auto stride = strides[dim];
       for (int i = 0; i < size; i++) {
         loop_inner(dim + 1, offset + i * stride);
       }
     } else {
-      int size = shape[dim];
+      auto size = shape[dim];
-      size_t stride = strides[dim];
+      auto stride = strides[dim];
       for (int i = 0; i < size; i++) {
         callback(offset + i * stride);
       }
@@ -190,46 +198,114 @@ void nd_loop(
 void Reduce::eval(const std::vector<array>& inputs, array& out) {
   assert(inputs.size() == 1);
   auto& in = inputs[0];
-  switch (in.dtype()) {
+  switch (reduce_type_) {
-    case bool_:
+    case Reduce::And:
-      reduce_dispatch_out<bool>(in, out, reduce_type_, axes_);
+    case Reduce::Or: {
+      switch (in.dtype()) {
+        case bool_:
+        case uint8:
+        case int8:
+          reduce_dispatch_and_or<int8_t>(in, out, reduce_type_, axes_);
+          break;
+        case int16:
+        case uint16:
+        case float16:
+        case bfloat16:
+          reduce_dispatch_and_or<int16_t>(in, out, reduce_type_, axes_);
+          break;
+        case uint32:
+        case int32:
+        case float32:
+          reduce_dispatch_and_or<int32_t>(in, out, reduce_type_, axes_);
+          break;
+        case uint64:
+        case int64:
+        case complex64:
+          reduce_dispatch_and_or<int64_t>(in, out, reduce_type_, axes_);
+          break;
+      }
       break;
-    case uint8:
+    }
-      reduce_dispatch_out<uint8_t>(in, out, reduce_type_, axes_);
+    case Reduce::Sum:
+    case Reduce::Prod: {
+      switch (in.dtype()) {
+        case bool_:
+        case uint8:
+        case int8:
+          reduce_dispatch_sum_prod<int8_t>(in, out, reduce_type_, axes_);
+          break;
+        case int16:
+        case uint16:
+          reduce_dispatch_sum_prod<int16_t>(in, out, reduce_type_, axes_);
+          break;
+        case int32:
+        case uint32:
+          reduce_dispatch_sum_prod<int32_t>(in, out, reduce_type_, axes_);
+          break;
+        case int64:
+        case uint64:
+          reduce_dispatch_sum_prod<int64_t>(in, out, reduce_type_, axes_);
+          break;
+        case float16:
+          reduce_dispatch_sum_prod<float16_t>(in, out, reduce_type_, axes_);
+          break;
+        case bfloat16:
+          reduce_dispatch_sum_prod<bfloat16_t>(in, out, reduce_type_, axes_);
+          break;
+        case float32:
+          reduce_dispatch_sum_prod<float>(in, out, reduce_type_, axes_);
+          break;
+        case complex64:
+          reduce_dispatch_sum_prod<complex64_t>(in, out, reduce_type_, axes_);
+          break;
+      }
       break;
-    case uint16:
+    }
-      reduce_dispatch_out<uint16_t>(in, out, reduce_type_, axes_);
+    case Reduce::Max:
-      break;
+    case Reduce::Min: {
-    case uint32:
+      switch (in.dtype()) {
-      reduce_dispatch_out<uint32_t>(in, out, reduce_type_, axes_);
+        case bool_:
-      break;
+          reduce_dispatch_min_max<bool>(in, out, reduce_type_, axes_);
-    case uint64:
+          break;
-      reduce_dispatch_out<uint64_t>(in, out, reduce_type_, axes_);
+        case uint8:
-      break;
+          reduce_dispatch_min_max<uint8_t>(in, out, reduce_type_, axes_);
-    case int8:
+          break;
-      reduce_dispatch_out<uint8_t>(in, out, reduce_type_, axes_);
+        case uint16:
-      break;
+          reduce_dispatch_min_max<uint16_t>(in, out, reduce_type_, axes_);
-    case int16:
+          break;
-      reduce_dispatch_out<uint16_t>(in, out, reduce_type_, axes_);
+        case uint32:
-      break;
+          reduce_dispatch_min_max<uint32_t>(in, out, reduce_type_, axes_);
-    case int32:
+          break;
-      reduce_dispatch_out<int32_t>(in, out, reduce_type_, axes_);
+        case uint64:
-      break;
+          reduce_dispatch_min_max<uint64_t>(in, out, reduce_type_, axes_);
-    case int64:
+          break;
-      reduce_dispatch_out<int64_t>(in, out, reduce_type_, axes_);
+        case int8:
-      break;
+          reduce_dispatch_min_max<uint8_t>(in, out, reduce_type_, axes_);
-    case float16:
+          break;
-      reduce_dispatch_out<float16_t>(in, out, reduce_type_, axes_);
+        case int16:
-      break;
+          reduce_dispatch_min_max<uint16_t>(in, out, reduce_type_, axes_);
-    case float32:
+          break;
-      reduce_dispatch_out<float>(in, out, reduce_type_, axes_);
+        case int32:
-      break;
+          reduce_dispatch_min_max<int32_t>(in, out, reduce_type_, axes_);
-    case bfloat16:
+          break;
-      reduce_dispatch_out<bfloat16_t>(in, out, reduce_type_, axes_);
+        case int64:
-      break;
+          reduce_dispatch_min_max<int64_t>(in, out, reduce_type_, axes_);
-    case complex64:
+          break;
-      reduce_dispatch_out<complex64_t>(in, out, reduce_type_, axes_);
+        case float16:
+          reduce_dispatch_min_max<float16_t>(in, out, reduce_type_, axes_);
+          break;
+        case float32:
+          reduce_dispatch_min_max<float>(in, out, reduce_type_, axes_);
+          break;
+        case bfloat16:
+          reduce_dispatch_min_max<bfloat16_t>(in, out, reduce_type_, axes_);
+          break;
+        case complex64:
+          reduce_dispatch_min_max<complex64_t>(in, out, reduce_type_, axes_);
+          break;
+      }
       break;
+    }
   }
 }
 

mlx/backend/common/reduce.h

@@ -38,13 +38,10 @@ enum ReductionOpType {
 
 struct ReductionPlan {
   ReductionOpType type;
-  std::vector<int> shape;
+  Shape shape;
-  std::vector<size_t> strides;
+  Strides strides;
 
-  ReductionPlan(
+  ReductionPlan(ReductionOpType type_, Shape shape_, Strides strides_)
-      ReductionOpType type_,
-      std::vector<int> shape_,
-      std::vector<size_t> strides_)
       : type(type_), shape(std::move(shape_)), strides(std::move(strides_)) {}
   ReductionPlan(ReductionOpType type_) : type(type_) {}
 };
@@ -55,10 +52,10 @@ ReductionPlan get_reduction_plan(const array& x, const std::vector<int>& axes);
 // Should this be in utils?
 void nd_loop(
     std::function<void(int)> callback,
-    const std::vector<int>& shape,
+    const Shape& shape,
-    const std::vector<size_t>& strides);
+    const Strides& strides);
 
-std::pair<std::vector<int>, std::vector<size_t>> shapes_without_reduction_axes(
+std::pair<Shape, Strides> shapes_without_reduction_axes(
     const array& x,
     const std::vector<int>& axes);
 
@@ -113,9 +110,6 @@ void reduction_op(
     return;
   }
 
-  std::vector<int> shape;
-  std::vector<size_t> strides;
-
   if (plan.type == ContiguousReduce && plan.shape.size() == 1) {
     int reduction_size = plan.shape[0];
     const T* x_ptr = x.data<T>();
@@ -135,7 +129,7 @@ void reduction_op(
     U* out_ptr = out.data<U>();
     // Unrolling the following loop (and implementing it in order for
     // ContiguousReduce) should hold extra performance boost.
-    std::tie(shape, strides) = shapes_without_reduction_axes(x, axes);
+    auto [shape, strides] = shapes_without_reduction_axes(x, axes);
     if (plan.shape.size() == 0) {
       for (int i = 0; i < out.size(); i++, out_ptr++) {
         int offset = elem_to_loc(i, shape, strides);
@@ -181,7 +175,7 @@ void reduction_op(
     plan.strides.pop_back();
     const T* x_ptr = x.data<T>();
     U* out_ptr = out.data<U>();
-    std::tie(shape, strides) = shapes_without_reduction_axes(x, axes);
+    auto [shape, strides] = shapes_without_reduction_axes(x, axes);
     if (plan.shape.size() == 0) {
       for (int i = 0; i < out.size(); i += reduction_stride) {
         int offset = elem_to_loc(i, shape, strides);
@@ -211,7 +205,7 @@ void reduction_op(
   if (plan.type == GeneralReduce) {
     const T* x_ptr = x.data<T>();
     U* out_ptr = out.data<U>();
-    std::tie(shape, strides) = shapes_without_reduction_axes(x, axes);
+    auto [shape, strides] = shapes_without_reduction_axes(x, axes);
     for (int i = 0; i < out.size(); i++, out_ptr++) {
       int offset = elem_to_loc(i, shape, strides);
       U val = init;

mlx/backend/common/reduce_utils.cpp

@@ -4,11 +4,11 @@
 
 namespace mlx::core {
 
-std::pair<std::vector<int>, std::vector<size_t>> shapes_without_reduction_axes(
+std::pair<Shape, Strides> shapes_without_reduction_axes(
     const array& x,
     const std::vector<int>& axes) {
-  std::vector<int> shape = x.shape();
+  auto shape = x.shape();
-  std::vector<size_t> strides = x.strides();
+  auto strides = x.strides();
 
   for (int i = axes.size() - 1; i >= 0; i--) {
     int a = axes[i];
@@ -29,8 +29,8 @@ ReductionPlan get_reduction_plan(const array& x, const std::vector<int>& axes) {
   // Row contiguous input so the output is row contiguous
   if (x.flags().row_contiguous) {
     // Merge consecutive axes
-    std::vector<int> shape = {x.shape(axes[0])};
+    Shape shape = {x.shape(axes[0])};
-    std::vector<size_t> strides = {x.strides()[axes[0]]};
+    Strides strides = {x.strides()[axes[0]]};
     for (int i = 1; i < axes.size(); i++) {
       if (axes[i] - 1 == axes[i - 1] && x.shape(axes[i]) > 1) {
         shape.back() *= x.shape(axes[i]);
@@ -69,7 +69,7 @@ ReductionPlan get_reduction_plan(const array& x, const std::vector<int>& axes) {
 
   // Sort reduction axes by stride in order to merge them and figure out if we
   // have a contiguous reduction.
-  std::vector<std::pair<int, size_t>> reductions;
+  std::vector<std::pair<int, int64_t>> reductions;
   for (auto a : axes) {
     if (x.shape(a) > 1) {
       reductions.push_back(std::make_pair(x.shape(a), x.strides()[a]));
@@ -93,8 +93,8 @@ ReductionPlan get_reduction_plan(const array& x, const std::vector<int>& axes) {
     }
   }
 
-  std::vector<int> shape;
+  Shape shape;
-  std::vector<size_t> strides;
+  Strides strides;
   for (auto r : reductions) {
     shape.push_back(r.first);
     strides.push_back(r.second);
@@ -109,15 +109,15 @@ ReductionPlan get_reduction_plan(const array& x, const std::vector<int>& axes) {
   // Delegate to the general strided reduction op if the axes after
   // strides.back() are contiguous.
   if (strides.back() > 1) {
-    int size = 1;
+    int64_t size = 1;
     bool have_expand = false;
     for (int i = x.ndim() - 1; i >= 0; i--) {
       if (axes.back() == i) {
         continue;
       }
 
-      size_t stride_i = x.strides()[i];
+      auto stride_i = x.strides()[i];
-      int shape_i = x.shape(i);
+      auto shape_i = x.shape(i);
       if (stride_i == 0) {
         if (shape_i == 1) {
           continue;

mlx/backend/common/slicing.cpp

@@ -4,24 +4,22 @@
 
 namespace mlx::core {
 
-std::tuple<bool, int64_t, std::vector<int64_t>> prepare_slice(
+std::tuple<int64_t, Strides> prepare_slice(
     const array& in,
-    const std::vector<int>& start_indices,
+    const Shape& start_indices,
-    const std::vector<int>& strides) {
+    const Shape& strides) {
   int64_t data_offset = 0;
-  bool copy_needed = false;
+  Strides inp_strides(in.ndim(), 0);
-  std::vector<int64_t> inp_strides(in.ndim(), 0);
   for (int i = 0; i < in.ndim(); ++i) {
     data_offset += start_indices[i] * in.strides()[i];
     inp_strides[i] = in.strides()[i] * strides[i];
-    copy_needed |= strides[i] < 0;
   }
-  return std::make_tuple(copy_needed, data_offset, inp_strides);
+  return std::make_tuple(data_offset, inp_strides);
 }
 
 void shared_buffer_slice(
     const array& in,
-    const std::vector<size_t>& out_strides,
+    const Strides& out_strides,
     size_t data_offset,
     size_t data_size,
     array& out) {
@@ -34,7 +32,7 @@ void shared_buffer_slice(
   flags.col_contiguous = is_col_contiguous;
   flags.contiguous = (no_bsx_size == data_size);
 
-  out.copy_shared_buffer(in, out_strides, flags, data_size, data_offset);
+  move_or_copy(in, out, out_strides, flags, data_size, data_offset);
 }
 
 } // namespace mlx::core

mlx/backend/common/slicing.h

@@ -6,14 +6,14 @@
 
 namespace mlx::core {
 
-std::tuple<bool, int64_t, std::vector<int64_t>> prepare_slice(
+std::tuple<int64_t, Strides> prepare_slice(
     const array& in,
-    const std::vector<int>& start_indices,
+    const Shape& start_indices,
-    const std::vector<int>& strides);
+    const Shape& strides);
 
 void shared_buffer_slice(
     const array& in,
-    const std::vector<size_t>& out_strides,
+    const Strides& out_strides,
     size_t data_offset,
     size_t data_size,
     array& out);

mlx/backend/common/sort.cpp

@@ -25,7 +25,7 @@ struct StridedIterator {
   // Constructors
   StridedIterator() = default;
 
-  explicit StridedIterator(T* ptr, size_t stride, difference_type offset = 0)
+  explicit StridedIterator(T* ptr, int64_t stride, difference_type offset = 0)
       : ptr_(ptr + offset * stride), stride_(stride) {}
 
   explicit StridedIterator(array& arr, int axis, difference_type offset = 0)
@@ -99,7 +99,7 @@ struct StridedIterator {
   }
 
  private:
-  size_t stride_;
+  int64_t stride_;
   T* ptr_;
 };
 
@@ -120,11 +120,11 @@ void sort(const array& in, array& out, int axis) {
   auto remaining_strides = out.strides();
   remaining_strides.erase(remaining_strides.begin() + axis);
 
-  size_t axis_stride = out.strides()[axis];
+  auto axis_stride = out.strides()[axis];
-  int axis_size = out.shape(axis);
+  auto axis_size = out.shape(axis);
 
   // Perform sorting in place
-  ContiguousIterator<size_t> src_it(
+  ContiguousIterator src_it(
       remaining_shape, remaining_strides, remaining_shape.size());
   for (int i = 0; i < n_rows; i++) {
     T* data_ptr = out.data<T>() + src_it.loc;
@@ -158,14 +158,14 @@ void argsort(const array& in, array& out, int axis) {
   auto out_remaining_strides = out.strides();
   out_remaining_strides.erase(out_remaining_strides.begin() + axis);
 
-  size_t in_stride = in.strides()[axis];
+  auto in_stride = in.strides()[axis];
-  size_t out_stride = out.strides()[axis];
+  auto out_stride = out.strides()[axis];
-  int axis_size = in.shape(axis);
+  auto axis_size = in.shape(axis);
 
   // Perform sorting
-  ContiguousIterator<size_t> in_it(
+  ContiguousIterator in_it(
       in_remaining_shape, in_remaining_strides, in_remaining_shape.size());
-  ContiguousIterator<size_t> out_it(
+  ContiguousIterator out_it(
       out_remaining_shape, out_remaining_strides, out_remaining_shape.size());
   for (int i = 0; i < n_rows; i++) {
     const T* data_ptr = in.data<T>() + in_it.loc;
@@ -208,13 +208,13 @@ void partition(const array& in, array& out, int axis, int kth) {
   auto remaining_strides = in.strides();
   remaining_strides.erase(remaining_strides.begin() + axis);
 
-  size_t axis_stride = in.strides()[axis];
+  auto axis_stride = in.strides()[axis];
   int axis_size = in.shape(axis);
 
   kth = kth < 0 ? kth + axis_size : kth;
 
   // Perform partition in place
-  ContiguousIterator<size_t> src_it(
+  ContiguousIterator src_it(
       remaining_shape, remaining_strides, remaining_shape.size());
   for (int i = 0; i < n_rows; i++) {
     T* data_ptr = out.data<T>() + src_it.loc;
@@ -249,16 +249,16 @@ void argpartition(const array& in, array& out, int axis, int kth) {
   auto out_remaining_strides = out.strides();
   out_remaining_strides.erase(out_remaining_strides.begin() + axis);
 
-  size_t in_stride = in.strides()[axis];
+  auto in_stride = in.strides()[axis];
-  size_t out_stride = out.strides()[axis];
+  auto out_stride = out.strides()[axis];
-  int axis_size = in.shape(axis);
+  auto axis_size = in.shape(axis);
 
   kth = kth < 0 ? kth + axis_size : kth;
 
   // Perform partition
-  ContiguousIterator<size_t> in_it(
+  ContiguousIterator in_it(
       in_remaining_shape, in_remaining_strides, in_remaining_shape.size());
-  ContiguousIterator<size_t> out_it(
+  ContiguousIterator out_it(
       out_remaining_shape, out_remaining_strides, out_remaining_shape.size());
   for (int i = 0; i < n_rows; i++) {
     const T* data_ptr = in.data<T>() + in_it.loc;

mlx/backend/common/ternary.h

@@ -78,11 +78,11 @@ void ternary_op_dims(
     const T3* c,
     U* out,
     Op op,
-    const std::vector<int>& shape,
+    const Shape& shape,
-    const std::vector<size_t>& a_strides,
+    const Strides& a_strides,
-    const std::vector<size_t>& b_strides,
+    const Strides& b_strides,
-    const std::vector<size_t>& c_strides,
+    const Strides& c_strides,
-    const std::vector<size_t>& out_strides,
+    const Strides& out_strides,
     int axis) {
   auto stride_a = a_strides[axis];
   auto stride_b = b_strides[axis];
@@ -164,10 +164,10 @@ void ternary_op_dispatch_dims(
       return;
   }
 
-  ContiguousIterator<size_t> a_it(shape, a_strides, ndim - 2);
+  ContiguousIterator a_it(shape, a_strides, ndim - 2);
-  ContiguousIterator<size_t> b_it(shape, b_strides, ndim - 2);
+  ContiguousIterator b_it(shape, b_strides, ndim - 2);
-  ContiguousIterator<size_t> c_it(shape, c_strides, ndim - 2);
+  ContiguousIterator c_it(shape, c_strides, ndim - 2);
-  size_t stride = out_strides[ndim - 3];
+  auto stride = out_strides[ndim - 3];
   for (size_t elem = 0; elem < a.size(); elem += stride) {
     ternary_op_dims<T1, T2, T3, U, Op, 2>(
         a_ptr + a_it.loc,

mlx/backend/common/utils.cpp

@@ -4,15 +4,35 @@
 
 namespace mlx::core {
 
-template <typename StrideT>
+void move_or_copy(const array& in, array& out) {
-std::tuple<std::vector<int>, std::vector<std::vector<StrideT>>>
+  if (in.is_donatable()) {
-collapse_contiguous_dims_impl(
+    out.move_shared_buffer(in);
-    const std::vector<int>& shape,
+  } else {
-    const std::vector<std::vector<StrideT>>& strides,
+    out.copy_shared_buffer(in);
-    StrideT size_cap) {
+  }
+}
+
+void move_or_copy(
+    const array& in,
+    array& out,
+    const Strides& strides,
+    array::Flags flags,
+    size_t data_size,
+    size_t offset /* = 0 */) {
+  if (in.is_donatable()) {
+    out.move_shared_buffer(in, strides, flags, data_size, offset);
+  } else {
+    out.copy_shared_buffer(in, strides, flags, data_size, offset);
+  }
+}
+
+std::tuple<Shape, std::vector<Strides>> collapse_contiguous_dims(
+    const Shape& shape,
+    const std::vector<Strides>& strides,
+    int64_t size_cap) {
   // Make a vector that has axes separated with -1. Collapse all axes between
   // -1.
-  std::vector<int> to_collapse;
+  Shape to_collapse;
   if (shape.size() > 0) {
     if (shape[0] != 1) {
       to_collapse.push_back(0);
@@ -21,7 +41,7 @@ collapse_contiguous_dims_impl(
     for (int i = 1; i < shape.size(); i++) {
       bool contiguous = true;
       size *= shape[i];
-      for (const std::vector<StrideT>& st : strides) {
+      for (const auto& st : strides) {
         if (st[i] * shape[i] != st[i - 1] || size > size_cap) {
           contiguous = false;
           size = shape[i];
@@ -38,8 +58,8 @@ collapse_contiguous_dims_impl(
     to_collapse.push_back(-1);
   }
 
-  std::vector<int> out_shape;
+  Shape out_shape;
-  std::vector<std::vector<StrideT>> out_strides(strides.size());
+  std::vector<Strides> out_strides(strides.size());
   for (int i = 0;;) {
     while (i < to_collapse.size() && to_collapse[i] == -1) {
       ++i;
@@ -54,7 +74,7 @@ collapse_contiguous_dims_impl(
     }
     out_shape.push_back(current_shape);
     for (int j = 0; j < strides.size(); j++) {
-      const std::vector<StrideT>& st = strides[j];
+      const auto& st = strides[j];
       out_strides[j].push_back(st[to_collapse[k - 1]]);
     }
     i = k + 1;
@@ -69,29 +89,12 @@ collapse_contiguous_dims_impl(
   return std::make_tuple(out_shape, out_strides);
 }
 
-std::tuple<std::vector<int>, std::vector<std::vector<int64_t>>>
+std::pair<Shape, Strides> collapse_contiguous_dims(
-collapse_contiguous_dims(
+    const Shape& shape,
-    const std::vector<int>& shape,
+    const Strides& strides,
-    const std::vector<std::vector<int64_t>>& strides,
+    int64_t size_cap) {
-    int64_t size_cap /* = std::numeric_limits<int32_t>::max() */) {
+  Shape collapsed_shape;
-  return collapse_contiguous_dims_impl(shape, strides, size_cap);
+  Strides collapsed_strides;
-}
-
-std::tuple<std::vector<int>, std::vector<std::vector<size_t>>>
-collapse_contiguous_dims(
-    const std::vector<int>& shape,
-    const std::vector<std::vector<size_t>>& strides,
-    size_t size_cap /* = std::numeric_limits<int32>::max() */) {
-  return collapse_contiguous_dims_impl(shape, strides, size_cap);
-}
-
-template <typename StrideT>
-std::pair<std::vector<int>, std::vector<StrideT>> collapse_contiguous_dims_impl(
-    const std::vector<int>& shape,
-    const std::vector<StrideT>& strides,
-    StrideT size_cap) {
-  std::vector<int> collapsed_shape;
-  std::vector<StrideT> collapsed_strides;
 
   if (shape.size() > 0) {
     collapsed_shape.push_back(shape[0]);
@@ -101,7 +104,7 @@ std::pair<std::vector<int>, std::vector<StrideT>> collapse_contiguous_dims_impl(
         continue;
       } else if (
           strides[i] * shape[i] != collapsed_strides.back() ||
-          collapsed_shape.back() * static_cast<StrideT>(shape[i]) > size_cap) {
+          collapsed_shape.back() * static_cast<int64_t>(shape[i]) > size_cap) {
         collapsed_shape.push_back(shape[i]);
         collapsed_strides.push_back(strides[i]);
       } else {
@@ -114,25 +117,10 @@ std::pair<std::vector<int>, std::vector<StrideT>> collapse_contiguous_dims_impl(
   return std::make_pair(collapsed_shape, collapsed_strides);
 }
 
-std::pair<std::vector<int>, std::vector<int64_t>> collapse_contiguous_dims(
+std::pair<Shape, Strides> collapse_contiguous_dims(
-    const std::vector<int>& shape,
-    const std::vector<int64_t>& strides,
-    int64_t size_cap /* = std::numeric_limits<int32_t>::max() */) {
-  return collapse_contiguous_dims_impl<int64_t>(shape, strides, size_cap);
-}
-
-std::pair<std::vector<int>, std::vector<size_t>> collapse_contiguous_dims(
-    const std::vector<int>& shape,
-    const std::vector<size_t>& strides,
-    size_t size_cap /* = std::numeric_limits<int32_t>::max() */) {
-  return collapse_contiguous_dims_impl<size_t>(shape, strides, size_cap);
-}
-
-std::pair<std::vector<int>, std::vector<size_t>> collapse_contiguous_dims(
     const array& a,
-    size_t size_cap /* = std::numeric_limits<int32_t>::max()*/) {
+    int64_t size_cap /* = std::numeric_limits<int32_t>::max()*/) {
-  return collapse_contiguous_dims_impl<size_t>(
+  return collapse_contiguous_dims(a.shape(), a.strides(), size_cap);
-      a.shape(), a.strides(), size_cap);
 }
 
 } // namespace mlx::core

mlx/backend/common/utils.h

@@ -8,12 +8,9 @@
 
 namespace mlx::core {
 
-template <typename StrideT>
+inline int64_t
-inline StrideT elem_to_loc(
+elem_to_loc(int elem, const Shape& shape, const Strides& strides) {
-    int elem,
+  int64_t loc = 0;
-    const std::vector<int>& shape,
-    const std::vector<StrideT>& strides) {
-  StrideT loc = 0;
   for (int i = shape.size() - 1; i >= 0; --i) {
     auto q_and_r = ldiv(elem, shape[i]);
     loc += q_and_r.rem * strides[i];
@@ -22,16 +19,15 @@ inline StrideT elem_to_loc(
   return loc;
 }
 
-inline size_t elem_to_loc(int elem, const array& a) {
+inline int64_t elem_to_loc(int elem, const array& a) {
   if (a.flags().row_contiguous) {
     return elem;
   }
   return elem_to_loc(elem, a.shape(), a.strides());
 }
 
-template <typename StrideT>
+inline Strides make_contiguous_strides(const Shape& shape) {
-std::vector<StrideT> make_contiguous_strides(const std::vector<int>& shape) {
+  Strides strides(shape.size(), 1);
-  std::vector<StrideT> strides(shape.size(), 1);
   for (int i = shape.size() - 1; i > 0; i--) {
     strides[i - 1] = strides[i] * shape[i];
   }
@@ -44,22 +40,15 @@ std::vector<StrideT> make_contiguous_strides(const std::vector<int>& shape) {
 //
 // When multiple arrays are passed they should all have the same shape. The
 // collapsed axes are also the same so one shape is returned.
-std::tuple<std::vector<int>, std::vector<std::vector<int64_t>>>
+std::tuple<Shape, std::vector<Strides>> collapse_contiguous_dims(
-collapse_contiguous_dims(
+    const Shape& shape,
-    const std::vector<int>& shape,
+    const std::vector<Strides>& strides,
-    const std::vector<std::vector<int64_t>>& strides,
     int64_t size_cap = std::numeric_limits<int32_t>::max());
-std::tuple<std::vector<int>, std::vector<std::vector<size_t>>>
-collapse_contiguous_dims(
-    const std::vector<int>& shape,
-    const std::vector<std::vector<size_t>>& strides,
-    size_t size_cap = std::numeric_limits<int32_t>::max());
 
-inline std::tuple<std::vector<int>, std::vector<std::vector<size_t>>>
+inline std::tuple<Shape, std::vector<Strides>> collapse_contiguous_dims(
-collapse_contiguous_dims(
     const std::vector<array>& xs,
     size_t size_cap = std::numeric_limits<int32_t>::max()) {
-  std::vector<std::vector<size_t>> strides;
+  std::vector<Strides> strides;
   for (auto& x : xs) {
     strides.emplace_back(x.strides());
   }
@@ -73,19 +62,14 @@ inline auto collapse_contiguous_dims(Arrays&&... xs) {
 }
 
 // The single array version of the above.
-std::pair<std::vector<int>, std::vector<int64_t>> collapse_contiguous_dims(
+std::pair<Shape, Strides> collapse_contiguous_dims(
-    const std::vector<int>& shape,
+    const Shape& shape,
-    const std::vector<int64_t>& strides,
+    const Strides& strides,
     int64_t size_cap = std::numeric_limits<int32_t>::max());
-std::pair<std::vector<int>, std::vector<size_t>> collapse_contiguous_dims(
+std::pair<Shape, Strides> collapse_contiguous_dims(
-    const std::vector<int>& shape,
-    const std::vector<size_t>& strides,
-    size_t size_cap = std::numeric_limits<int32_t>::max());
-std::pair<std::vector<int>, std::vector<size_t>> collapse_contiguous_dims(
     const array& a,
-    size_t size_cap = std::numeric_limits<int32_t>::max());
+    int64_t size_cap = std::numeric_limits<int32_t>::max());
 
-template <typename StrideT>
 struct ContiguousIterator {
   inline void step() {
     int dims = shape_.size();
@@ -102,7 +86,7 @@ struct ContiguousIterator {
     loc += strides_[i];
   }
 
-  void seek(StrideT n) {
+  void seek(int64_t n) {
     loc = 0;
     for (int i = shape_.size() - 1; i >= 0; --i) {
       auto q_and_r = ldiv(n, shape_[i]);
@@ -128,32 +112,29 @@ struct ContiguousIterator {
   }
 
   explicit ContiguousIterator(
-      const std::vector<int>& shape,
+      const Shape& shape,
-      const std::vector<StrideT>& strides,
+      const Strides& strides,
       int dims)
       : shape_(shape.begin(), shape.begin() + dims),
         strides_(strides.begin(), strides.begin() + dims) {
     if (!shape_.empty()) {
       std::tie(shape_, strides_) = collapse_contiguous_dims(shape_, strides_);
-      pos_ = std::vector<int>(shape_.size(), 0);
+      pos_ = Shape(shape_.size(), 0);
     }
   }
 
-  StrideT loc{0};
+  int64_t loc{0};
 
  private:
-  std::vector<int> shape_;
+  Shape shape_;
-  std::vector<StrideT> strides_;
+  Strides strides_;
-  std::vector<int> pos_;
+  Shape pos_;
 };
 
-template <typename StrideT>
+inline auto check_contiguity(const Shape& shape, const Strides& strides) {
-inline auto check_contiguity(
-    const std::vector<int>& shape,
-    const std::vector<StrideT>& strides) {
   size_t no_broadcast_data_size = 1;
-  size_t f_stride = 1;
+  int64_t f_stride = 1;
-  size_t b_stride = 1;
+  int64_t b_stride = 1;
   bool is_row_contiguous = true;
   bool is_col_contiguous = true;
 
@@ -178,4 +159,13 @@ inline bool is_donatable(const array& in, const array& out) {
       in.buffer_size() <= out.nbytes() + donation_extra;
 }
 
+void move_or_copy(const array& in, array& out);
+void move_or_copy(
+    const array& in,
+    array& out,
+    const Strides& strides,
+    array::Flags flags,
+    size_t data_size,
+    size_t offset = 0);
+
 } // namespace mlx::core

mlx/backend/metal/CMakeLists.txt

@@ -14,14 +14,21 @@ function(make_jit_source SRC_FILE)
     COMMAND
       /bin/bash ${CMAKE_CURRENT_SOURCE_DIR}/make_compiled_preamble.sh
       ${CMAKE_CURRENT_BINARY_DIR}/jit ${CMAKE_C_COMPILER} ${PROJECT_SOURCE_DIR}
-      ${SRC_FILE} "-DMLX_METAL_VERSION=${MLX_METAL_VERSION}"
+      ${SRC_FILE}
     DEPENDS make_compiled_preamble.sh kernels/${SRC_FILE}.h ${ARGN})
   add_custom_target(${SRC_NAME} DEPENDS jit/${SRC_NAME}.cpp)
   add_dependencies(mlx ${SRC_NAME})
   target_sources(mlx PRIVATE ${CMAKE_CURRENT_BINARY_DIR}/jit/${SRC_NAME}.cpp)
 endfunction(make_jit_source)
 
-make_jit_source(utils kernels/bf16.h kernels/complex.h kernels/defines.h)
+make_jit_source(
+  utils
+  kernels/jit/bf16.h
+  kernels/metal_3_0/bf16.h
+  kernels/metal_3_1/bf16.h
+  kernels/bf16_math.h
+  kernels/complex.h
+  kernels/defines.h)
 make_jit_source(unary_ops kernels/erf.h kernels/expm1f.h)
 make_jit_source(binary_ops)
 make_jit_source(ternary_ops)

mlx/backend/metal/allocator.cpp

@@ -30,7 +30,7 @@ BufferCache::BufferCache(MTL::Device* device)
     : device_(device), head_(nullptr), tail_(nullptr), pool_size_(0) {}
 
 BufferCache::~BufferCache() {
-  auto thread_pool = metal::new_scoped_memory_pool();
+  auto pool = metal::new_scoped_memory_pool();
   clear();
 }
 
@@ -155,11 +155,13 @@ MetalAllocator::MetalAllocator()
 }
 
 size_t MetalAllocator::set_cache_limit(size_t limit) {
+  std::unique_lock lk(mutex_);
   std::swap(limit, max_pool_size_);
   return limit;
 };
 
 size_t MetalAllocator::set_memory_limit(size_t limit, bool relaxed) {
+  std::unique_lock lk(mutex_);
   std::swap(limit, block_limit_);
   relaxed_ = relaxed;
   gc_limit_ = std::min(
@@ -169,6 +171,7 @@ size_t MetalAllocator::set_memory_limit(size_t limit, bool relaxed) {
 };
 
 size_t MetalAllocator::set_wired_limit(size_t limit) {
+  std::unique_lock lk(mutex_);
   std::swap(limit, wired_limit_);
   residency_set_.resize(wired_limit_);
   return limit;
@@ -205,7 +208,7 @@ Buffer MetalAllocator::malloc(size_t size, bool allow_swap /* = false */) {
       return Buffer{nullptr};
     }
 
-    auto thread_pool = metal::new_scoped_memory_pool();
+    auto pool = metal::new_scoped_memory_pool();
 
     // If we have a lot of memory pressure or are over the maximum cache size,
     // try to reclaim memory from the cache
@@ -226,7 +229,7 @@ Buffer MetalAllocator::malloc(size_t size, bool allow_swap /* = false */) {
 
   // Maintain the cache below the requested limit
   if (get_cache_memory() >= max_pool_size_) {
-    auto thread_pool = metal::new_scoped_memory_pool();
+    auto pool = metal::new_scoped_memory_pool();
     buffer_cache_.release_cached_buffers(get_cache_memory() - max_pool_size_);
   }
 
@@ -237,11 +240,15 @@ Buffer MetalAllocator::malloc(size_t size, bool allow_swap /* = false */) {
 
 void MetalAllocator::clear_cache() {
   std::unique_lock lk(mutex_);
+  auto pool = metal::new_scoped_memory_pool();
   buffer_cache_.clear();
 }
 
 void MetalAllocator::free(Buffer buffer) {
   auto buf = static_cast<MTL::Buffer*>(buffer.ptr());
+  if (buf == nullptr) {
+    return;
+  }
   std::unique_lock lk(mutex_);
   residency_set_.erase(buf);
   active_memory_ -= buf->length();
@@ -249,7 +256,7 @@ void MetalAllocator::free(Buffer buffer) {
     buffer_cache_.recycle_to_cache(buf);
   } else {
     lk.unlock();
-    auto thread_pool = metal::new_scoped_memory_pool();
+    auto pool = metal::new_scoped_memory_pool();
     buf->release();
   }
 }

mlx/backend/metal/binary.cpp

@@ -22,37 +22,37 @@ std::string get_kernel_name(
     BinaryOpType bopt,
     const std::string& op,
     const array& a,
-    bool use_2d,
+    bool large,
     int ndim,
     int work_per_thread) {
-  std::ostringstream kname;
+  std::string kname;
   switch (bopt) {
     case BinaryOpType::ScalarScalar:
-      kname << "ss";
+      kname = "ss";
       break;
     case BinaryOpType::ScalarVector:
-      kname << (use_2d ? "sv2" : "sv");
+      kname = (large ? "sv2" : "sv");
       break;
     case BinaryOpType::VectorScalar:
-      kname << (use_2d ? "vs2" : "vs");
+      kname = (large ? "vs2" : "vs");
       break;
     case BinaryOpType::VectorVector:
-      kname << (use_2d ? "vv2" : "vv");
+      kname = (large ? "vv2" : "vv");
       break;
     case BinaryOpType::General:
-      kname << "g";
+      kname = "g";
       if (ndim <= 3) {
-        kname << ndim;
+        kname += std::to_string(ndim);
       } else {
-        kname << "n";
+        concatenate(kname, "n", std::to_string(work_per_thread));
-        if (work_per_thread > 1) {
+      }
-          kname << work_per_thread;
+      if (large) {
-        }
+        kname += "large";
       }
       break;
   }
-  kname << "_" << op << type_to_name(a);
+  concatenate(kname, "_", op, type_to_name(a));
-  return kname.str();
+  return kname;
 }
 
 void binary_op_gpu_inplace(
@@ -75,24 +75,30 @@ void binary_op_gpu_inplace(
       auto [shape, strides] = collapse_contiguous_dims(a, b, out);
       return std::make_tuple(shape, strides[0], strides[1], strides[2]);
     } else {
-      std::vector<size_t> e;
+      decltype(a.strides()) e{};
-      return std::make_tuple(std::vector<int>{}, e, e, e);
+      return std::make_tuple(decltype(a.shape()){}, e, e, e);
     }
   };
   auto [shape, strides_a, strides_b, strides_out] = maybe_collapse();
 
-  bool use_2d = out.data_size() > UINT32_MAX;
+  bool large = out.data_size() > UINT32_MAX;
   auto ndim = shape.size();
-  int work_per_thread = (bopt == BinaryOpType::General) ? 4 : 1;
+  int work_per_thread;
+  if (bopt == BinaryOpType::General) {
+    large |= (a.data_size() > UINT32_MAX || b.data_size() > UINT32_MAX);
+    work_per_thread = large ? 4 : 2;
+  } else {
+    work_per_thread = 1;
+  }
   std::string kernel_name =
-      get_kernel_name(bopt, op, a, use_2d, shape.size(), work_per_thread);
+      get_kernel_name(bopt, op, a, large, shape.size(), work_per_thread);
   auto& d = metal::device(s.device);
 
   auto kernel = outputs.size() == 2
       ? get_binary_two_kernel(d, kernel_name, a.dtype(), out.dtype(), op)
       : get_binary_kernel(d, kernel_name, a.dtype(), out.dtype(), op);
   auto& compute_encoder = d.get_command_encoder(s.index);
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
 
   // - If a is donated it goes to the first output
   // - If b is donated it goes to the first output if a was not donated
@@ -117,19 +123,15 @@ void binary_op_gpu_inplace(
     size_t rest = out.size() / (dim0 * dim1);
 
     if (ndim > 3) {
-      compute_encoder->setBytes(shape.data(), ndim * sizeof(int), arg_idx++);
+      compute_encoder.set_vector_bytes(shape, arg_idx++);
-      compute_encoder->setBytes(
+      compute_encoder.set_vector_bytes(strides_a, arg_idx++);
-          strides_a.data(), ndim * sizeof(size_t), arg_idx++);
+      compute_encoder.set_vector_bytes(strides_b, arg_idx++);
-      compute_encoder->setBytes(
+      compute_encoder.set_bytes<int>(ndim, arg_idx++);
-          strides_b.data(), ndim * sizeof(size_t), arg_idx++);
-      compute_encoder->setBytes(&ndim, sizeof(int), arg_idx++);
       dim0 = (dim0 + work_per_thread - 1) / work_per_thread;
     } else {
       // The shape is implicit in the grid for <= 3D
-      compute_encoder->setBytes(
+      compute_encoder.set_vector_bytes(strides_a, arg_idx++);
-          strides_a.data(), ndim * sizeof(size_t), arg_idx++);
+      compute_encoder.set_vector_bytes(strides_b, arg_idx++);
-      compute_encoder->setBytes(
-          strides_b.data(), ndim * sizeof(size_t), arg_idx++);
     }
 
     if (thread_group_size != 1024) {
@@ -137,7 +139,7 @@ void binary_op_gpu_inplace(
     }
     auto group_dims = get_block_dims(dim0, dim1, rest);
     MTL::Size grid_dims = MTL::Size(dim0, dim1, rest);
-    compute_encoder.dispatchThreads(grid_dims, group_dims);
+    compute_encoder.dispatch_threads(grid_dims, group_dims);
   } else {
     // Launch a 1D or 2D grid of threads
     size_t nthreads = out.data_size();
@@ -145,9 +147,9 @@ void binary_op_gpu_inplace(
       thread_group_size = nthreads;
     }
     MTL::Size group_dims = MTL::Size(thread_group_size, 1, 1);
-    MTL::Size grid_dims = use_2d ? get_2d_grid_dims(out.shape(), out.strides())
+    MTL::Size grid_dims = large ? get_2d_grid_dims(out.shape(), out.strides())
-                                 : MTL::Size(nthreads, 1, 1);
+                                : MTL::Size(nthreads, 1, 1);
-    compute_encoder.dispatchThreads(grid_dims, group_dims);
+    compute_encoder.dispatch_threads(grid_dims, group_dims);
   }
 }
 

mlx/backend/metal/compiled.cpp

@@ -1,5 +1,6 @@
 // Copyright © 2023-2024 Apple Inc.
-
+#include <fmt/format.h>
+#include <iostream> //TODO
 #include <sstream>
 
 #include "mlx/backend/common/compiled.h"
@@ -11,12 +12,12 @@
 #include "mlx/primitives.h"
 #include "mlx/utils.h"
 
+using namespace fmt::literals;
+
 namespace mlx::core {
 
-constexpr int WORK_PER_THREAD = 4;
-
 inline void build_kernel(
-    std::ostream& os,
+    std::string& os,
     const std::string& kernel_name,
     const std::vector<array>& inputs,
     const std::vector<array>& outputs,
@@ -41,8 +42,8 @@ inline void build_kernel(
   int cnt = 0;
 
   // Start the kernel
-  os << "[[host_name(\"" << kernel_name << "\")]]\n"
+  os += fmt::format(
-     << "[[kernel]] void " << kernel_name << "(\n";
+      "[[host_name(\"{0}\")]]\n[[kernel]] void {0}(\n", kernel_name);
 
   // Add the input arguments
   for (auto& x : inputs) {
@@ -54,51 +55,61 @@ inline void build_kernel(
     }
 
     // Scalars and contiguous need no strides
-    if (is_scalar(x) || contiguous) {
+    if (!is_scalar(x) && !contiguous) {
-      os << "    device const " << get_type_string(x.dtype()) << "* " << xname
-         << " [[buffer(" << cnt++ << ")]],\n";
-    } else {
       add_indices = true;
-      os << "    device const " << get_type_string(x.dtype()) << "* " << xname
-         << " [[buffer(" << cnt++ << ")]],\n";
     }
+    os += fmt::format(
+        "    device const {0}* {1} [[buffer({2})]],\n",
+        get_type_string(x.dtype()),
+        xname,
+        cnt++);
   }
 
   if (add_indices) {
-    os << "    constant const size_t* in_strides [[buffer(" << cnt++
+    os += fmt::format(
-       << ")]],\n";
+        "    constant const int64_t* in_strides [[buffer({0})]],\n", cnt++);
   }
 
   // Add the output arguments
   for (auto& x : outputs) {
-    os << "    device " << get_type_string(x.dtype()) << "* "
+    os += fmt::format(
-       << namer.get_name(x) << " [[buffer(" << cnt++ << ")]],\n";
+        "    device {0}* {1} [[buffer({2})]],\n",
+        get_type_string(x.dtype()),
+        namer.get_name(x),
+        cnt++);
   }
   // Add output strides and shape to extract the indices.
   if (!contiguous) {
-    os << "    constant const size_t* output_strides [[buffer(" << cnt++
+    os += fmt::format(
-       << ")]],\n"
+        "    constant const int64_t* output_strides [[buffer({0})]],\n", cnt++);
-       << "    constant const int* output_shape [[buffer(" << cnt++ << ")]],\n";
+    os += fmt::format(
+        "    constant const int* output_shape [[buffer({0})]],\n", cnt++);
   }
   if (dynamic_dims) {
-    os << "    constant const int& ndim [[buffer(" << cnt++ << ")]],\n";
+    os += fmt::format("    constant const int& ndim [[buffer({0})]],\n", cnt++);
   }
 
   // The thread index in the whole grid
-  os << "    uint3 pos [[thread_position_in_grid]],\n"
+  os += "    uint3 pos [[thread_position_in_grid]],\n";
-     << "    uint3 grid [[threads_per_grid]]) {\n";
+  os += "    uint3 grid [[threads_per_grid]]) {\n";
 
-  if (use_big_index) {
+  std::string idx_type = use_big_index ? "int64_t" : "uint";
+  if (contiguous && use_big_index) {
     // This is only used for contiguous kernels which don't have
     // a third grid dimension
-    os << "  size_t index = pos.x + grid.x * size_t(pos.y);\n";
+    os += "  int64_t index = pos.x + grid.x * int64_t(pos.y);\n";
   } else if (work_per_thread > 1) {
-    os << "  constexpr int N_ = " << std::to_string(work_per_thread) << ";\n"
+    os += fmt::format("  constexpr int N_ = {0};\n", work_per_thread);
-       << "  int xshape = output_shape["
+    os += fmt::format(
-       << (dynamic_dims ? "ndim - 1" : std::to_string(ndim - 1)) << "];\n"
+        "  int xshape = output_shape[{0}];\n",
-       << "  size_t index = N_ * pos.x + xshape * (pos.y + size_t(grid.y) * pos.z);\n";
+        dynamic_dims ? "ndim - 1" : std::to_string(ndim - 1));
+    os += fmt::format(
+        "  {0} index = N_ * pos.x + xshape * (pos.y + {0}(grid.y) * pos.z);\n",
+        idx_type);
   } else {
-    os << " size_t index = pos.x + grid.x * (pos.y + size_t(grid.y) * pos.z);\n";
+    os += fmt::format(
+        "  {0} index = pos.x + grid.x * (pos.y + {0}(grid.y) * pos.z);\n",
+        idx_type);
   }
 
   // Read constant / contiguous inputs in tmps
@@ -109,16 +120,19 @@ inline void build_kernel(
 
     if (is_constant(x)) {
       auto type_str = get_type_string(x.dtype());
-      os << "  auto tmp_" << xname << " = static_cast<"
+      std::ostringstream ss;
-         << get_type_string(x.dtype()) << ">(";
+      print_constant(ss, x);
-      print_constant(os, x);
+      os += fmt::format(
-      os << ");\n";
+          "  auto tmp_{0} = static_cast<{1}>({2});\n",
+          xname,
+          get_type_string(x.dtype()),
+          ss.str());
     } else if (is_scalar(x)) {
-      os << "  " << get_type_string(x.dtype()) << " tmp_" << xname << " = "
+      os += fmt::format(
-         << xname << "[0];\n";
+          "  {0} tmp_{1} = {1}[0];\n", get_type_string(x.dtype()), xname);
     } else if (contiguous) {
-      os << "  " << get_type_string(x.dtype()) << " tmp_" << xname << " = "
+      os += fmt::format(
-         << xname << "[index];\n";
+          "  {0} tmp_{1} = {1}[index];\n", get_type_string(x.dtype()), xname);
     } else {
       nc_inputs.push_back(x);
     }
@@ -127,83 +141,96 @@ inline void build_kernel(
   // Initialize the indices for non-contiguous inputs
   for (int i = 0; i < nc_inputs.size(); ++i) {
     auto& xname = namer.get_name(nc_inputs[i]);
+    os += fmt::format("  {0} index_{1} = ", idx_type, xname);
     if (ndim == 1) {
       int offset = i * ndim;
-      os << "  size_t index_" << xname << " = elem_to_loc_1(pos.x, "
+      os +=
-         << "in_strides[" << offset << "]);\n";
+          fmt::format("elem_to_loc_1<uint>(pos.x, in_strides[{0}]);\n", offset);
     } else if (ndim == 2) {
       int offset = i * ndim;
-      os << "  size_t index_" << xname << " = elem_to_loc_2({pos.x, pos.y}, "
+      os += fmt::format(
-         << "in_strides + " << offset << ");\n";
+          "elem_to_loc_2<{0}>({{pos.x, pos.y}}, in_strides + {1});\n",
+          idx_type,
+          offset);
     } else if (ndim == 3) {
       int offset = i * ndim;
-      os << "  size_t index_" << xname << " = elem_to_loc_3(pos, "
+      os += fmt::format(
-         << "in_strides + " << offset << ");\n";
+          "elem_to_loc_3<{0}>(pos, in_strides + {1});\n", idx_type, offset);
     } else if (!dynamic_dims) {
-      int offset = i * ndim;
+      int offset = (i + 1) * ndim;
-      os << "  size_t index_" << xname << " = N_ * pos.x * in_strides["
+      os += fmt::format(
-         << offset + ndim - 1 << "]"
+          "N_ * pos.x * {0}(in_strides[{1}]) + pos.y * {0}(in_strides[{2}]);\n",
-         << " + pos.y * in_strides[" << offset + ndim - 2 << "];\n";
+          idx_type,
+          offset - 1,
+          offset - 2);
     } else {
-      os << "  size_t index_" << xname << " = N_ * pos.x * in_strides[ndim * "
+      os += fmt::format(
-         << i << " + ndim - 1]"
+          "N_ * pos.x * {0}(in_strides[ndim * {1} + ndim - 1]) + pos.y * {0}(in_strides[ndim * {1} + ndim - 2]);\n",
-         << " + pos.y * in_strides[ndim * " << i << " + ndim - 2];\n";
+          idx_type,
+          i);
     }
   }
+
   if (!nc_inputs.empty() && (ndim > 3 || dynamic_dims)) {
-    os << "  uint zpos = pos.z;\n";
+    os += "  uint zpos = pos.z;\n";
     if (dynamic_dims) {
-      os << "  for (int d = ndim - 3; d >= 0; --d) {\n";
+      os += "  for (int d = ndim - 3; d >= 0; --d) {\n";
     } else {
-      os << "  for (int d = " << ndim - 3 << "; d >= 0; --d) {\n";
+      os += fmt::format("  for (int d = {0}; d >= 0; --d) {{\n", ndim - 3);
     }
-    os << "    uint l = zpos % output_shape[d];\n";
+    os += "    uint l = zpos % output_shape[d];\n";
     for (int i = 0; i < nc_inputs.size(); ++i) {
       auto& xname = namer.get_name(nc_inputs[i]);
-      os << "    index_" << xname << " += ";
+      os += fmt::format("    index_{0} += ", xname);
       if (dynamic_dims) {
-        os << "l * in_strides[" << i << " * ndim + d];\n";
+        os +=
+            fmt::format("l * {0}(in_strides[{1} * ndim + d]);\n", idx_type, i);
       } else {
-        os << "l * in_strides[" << i * ndim << " + d];\n";
+        os +=
+            fmt::format("l * {0}(in_strides[{1} + d]);\n", idx_type, i * ndim);
       }
     }
-    os << "    zpos /= output_shape[d];\n  }\n";
+    os += "    zpos /= output_shape[d];\n  }\n";
   }
 
   // Open per-thread loop
   if (work_per_thread > 1) {
-    os << "  for (int i = 0; i < N_ && (int(N_ * pos.x) + i) < xshape; ++i) {\n";
+    os +=
+        "  for (int i = 0; i < N_ && (int(N_ * pos.x) + i) < xshape; ++i) {\n";
   }
 
   // Read non-contiguous inputs into tmps
   for (int i = 0; i < nc_inputs.size(); ++i) {
     auto& x = nc_inputs[i];
     auto& xname = namer.get_name(x);
-    os << "  " << get_type_string(x.dtype()) << " tmp_" << xname << " = "
+    os += fmt::format(
-       << xname << "[index_" << xname << "];\n";
+        "  {0} tmp_{1} = {1}[index_{1}];\n", get_type_string(x.dtype()), xname);
   }
 
   // Actually write the computation
   for (auto& x : tape) {
-    os << "  " << get_type_string(x.dtype()) << " tmp_" << namer.get_name(x)
+    os += fmt::format(
-       << " = ";
+        "  {0} tmp_{1} = ", get_type_string(x.dtype()), namer.get_name(x));
     if (is_static_cast(x.primitive())) {
-      os << "static_cast<" << get_type_string(x.dtype()) << ">(tmp_"
+      os += fmt::format(
-         << namer.get_name(x.inputs()[0]) << ");\n";
+          "static_cast<{0}>(tmp_{1});\n",
+          get_type_string(x.dtype()),
+          namer.get_name(x.inputs()[0]));
     } else {
-      x.primitive().print(os);
+      std::ostringstream ss;
-      os << "()(";
+      x.primitive().print(ss);
+      os += ss.str();
+      os += "()(";
       for (int i = 0; i < x.inputs().size() - 1; i++) {
-        os << "tmp_" << namer.get_name(x.inputs()[i]) << ", ";
+        os += fmt::format("tmp_{0}, ", namer.get_name(x.inputs()[i]));
       }
-      os << "tmp_" << namer.get_name(x.inputs().back()) << ");\n";
+      os += fmt::format("tmp_{0});\n", namer.get_name(x.inputs().back()));
     }
   }
 
   // Write the outputs from tmps
   for (auto& x : outputs) {
-    os << "  " << namer.get_name(x) << "[index] = tmp_" << namer.get_name(x)
+    os += fmt::format("  {0}[index] = tmp_{0};\n", namer.get_name(x));
-       << ";\n";
   }
   // Increment indices and close per thread loop
   if (work_per_thread > 1) {
@@ -211,18 +238,18 @@ inline void build_kernel(
       auto& x = nc_inputs[i];
       auto& xname = namer.get_name(x);
       if (!dynamic_dims) {
-        os << "  index_" << xname << " += "
+        os += fmt::format(
-           << "in_strides[" << i * ndim + ndim - 1 << "];\n";
+            "  index_{0} += in_strides[{1}];\n", xname, i * ndim + ndim - 1);
       } else {
-        os << "  index_" << xname << " += "
+        os += fmt::format(
-           << "in_strides[" << i << " * ndim + ndim - 1];\n";
+            "  index_{0} += in_strides[{1} * ndim + ndim - 1];\n", xname, i);
       }
     }
-    os << "  index++;\n  }\n";
+    os += "  index++;\n  }\n";
   }
 
   // Finish the kernel
-  os << "}\n";
+  os += "}\n";
 
   if (cnt > 31) {
     std::ostringstream msg;
@@ -246,9 +273,9 @@ void Compiled::eval_gpu(
   auto& s = stream();
   auto& d = metal::device(s.device);
   auto lib = d.get_library(kernel_lib_, [&]() {
-    std::ostringstream kernel;
+    std::string kernel = metal::utils();
-    kernel << metal::utils() << metal::unary_ops() << metal::binary_ops()
+    concatenate(
-           << metal::ternary_ops();
+        kernel, metal::unary_ops(), metal::binary_ops(), metal::ternary_ops());
     build_kernel(
         kernel,
         kernel_lib_ + "_contiguous",
@@ -261,7 +288,7 @@ void Compiled::eval_gpu(
         /* dynamic_dims = */ false);
     build_kernel(
         kernel,
-        kernel_lib_ + "_contiguous_big",
+        kernel_lib_ + "_contiguous_large",
         inputs_,
         outputs_,
         tape_,
@@ -282,7 +309,21 @@ void Compiled::eval_gpu(
           /* ndim = */ i,
           /* dynamic_dims = */ false,
           /* use_big_index = */ false,
-          /* work_per_thread = */ i > 3 ? WORK_PER_THREAD : 1);
+          /* work_per_thread = */ i > 3 ? 2 : 1);
+      if (i > 1) {
+        build_kernel(
+            kernel,
+            kernel_lib_ + "_strided_" + std::to_string(i) + "_large",
+            inputs_,
+            outputs_,
+            tape_,
+            constant_ids_,
+            /* contiguous = */ false,
+            /* ndim = */ i,
+            /* dynamic_dims = */ false,
+            /* use_big_index = */ true,
+            /* work_per_thread = */ i > 3 ? 4 : 1);
+      }
     }
     build_kernel(
         kernel,
@@ -295,20 +336,32 @@ void Compiled::eval_gpu(
         /* ndim = */ 0,
         /* dynamic_dims = */ true,
         /* use_big_index = */ false,
-        /* work_per_thread = */ WORK_PER_THREAD);
+        /* work_per_thread = */ 2);
-    return kernel.str();
+    build_kernel(
+        kernel,
+        kernel_lib_ + "_strided_dynamic_large",
+        inputs_,
+        outputs_,
+        tape_,
+        constant_ids_,
+        /* contiguous = */ false,
+        /* ndim = */ 0,
+        /* dynamic_dims = */ true,
+        /* use_big_index = */ true,
+        /* work_per_thread = */ 4);
+    return kernel;
   });
 
   // Figure out which kernel we are using
   auto& output_shape = outputs[0].shape();
-  bool contiguous = compiled_check_contiguity(inputs, output_shape);
+  auto contiguous = compiled_check_contiguity(inputs, output_shape);
 
   // Collapse contiguous dims to route to a faster kernel if possible. Also
   // handle all broadcasting.
-  std::vector<std::vector<size_t>> initial_strides;
+  std::vector<Strides> initial_strides;
   initial_strides.push_back(outputs[0].strides());
-  std::vector<int> shape;
+  Shape shape;
-  std::vector<std::vector<size_t>> strides;
+  std::vector<Strides> strides;
   if (!contiguous) {
     for (int i = 0; i < inputs.size(); i++) {
       // Skip constants.
@@ -323,7 +376,7 @@ void Compiled::eval_gpu(
       }
 
       // Broadcast the inputs to the output shape.
-      std::vector<size_t> xstrides;
+      Strides xstrides;
       int j = 0;
       for (; j < output_shape.size() - x.ndim(); j++) {
         if (output_shape[j] == 1) {
@@ -349,13 +402,19 @@ void Compiled::eval_gpu(
         collapse_contiguous_dims(output_shape, initial_strides, INT32_MAX);
   }
 
-  bool use_2d = false;
+  bool large;
   if (contiguous) {
     size_t max_size = 0;
     for (auto& in : inputs) {
       max_size = std::max(max_size, in.data_size());
     }
-    use_2d = (max_size > UINT32_MAX);
+    large = (max_size > UINT32_MAX);
+  } else {
+    size_t max_size = 0;
+    for (auto& o : outputs) {
+      max_size = std::max(max_size, o.size());
+    }
+    large = (max_size > UINT32_MAX);
   }
 
   // Get the kernel from the lib
@@ -368,17 +427,18 @@ void Compiled::eval_gpu(
     } else {
       kernel_name += std::to_string(shape.size());
     }
-  } else if (use_2d) {
+  }
-    kernel_name += "_big";
+  if (large) {
+    kernel_name += "_large";
   }
   auto kernel = d.get_kernel(kernel_name, lib);
   auto& compute_encoder = d.get_command_encoder(s.index);
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
 
   // Put the inputs in
   int cnt = 0;
   int stride_idx = 1; // idx 0 is the output strides
-  std::vector<size_t> in_strides;
+  Strides in_strides;
   for (int i = 0; i < inputs.size(); i++) {
     if (constant_ids_.find(inputs_[i].id()) != constant_ids_.end()) {
       continue;
@@ -394,8 +454,7 @@ void Compiled::eval_gpu(
     }
   }
   if (!in_strides.empty()) {
-    compute_encoder->setBytes(
+    compute_encoder.set_vector_bytes(in_strides, cnt++);
-        in_strides.data(), in_strides.size() * sizeof(size_t), cnt++);
   }
 
   compiled_allocate_outputs(
@@ -408,14 +467,13 @@ void Compiled::eval_gpu(
 
   // Put the output shape and strides in
   if (!contiguous) {
-    compute_encoder->setBytes(
+    compute_encoder.set_vector_bytes(strides[0], cnt++);
-        strides[0].data(), strides[0].size() * sizeof(size_t), cnt++);
+    compute_encoder.set_vector_bytes(shape, cnt++);
-    compute_encoder->setBytes(shape.data(), shape.size() * sizeof(int), cnt++);
   }
 
   // Put the number of dims in if it is dynamic
   if (dynamic) {
-    compute_encoder->setBytes(&ndim, sizeof(int), cnt++);
+    compute_encoder.set_bytes(ndim, cnt++);
   }
 
   // Launch the kernel
@@ -424,15 +482,15 @@ void Compiled::eval_gpu(
     MTL::Size group_dims(
         std::min(nthreads, kernel->maxTotalThreadsPerThreadgroup()), 1, 1);
 
-    MTL::Size grid_dims = use_2d
+    MTL::Size grid_dims = large
         ? get_2d_grid_dims(outputs[0].shape(), outputs[0].strides())
         : MTL::Size(nthreads, 1, 1);
-    compute_encoder.dispatchThreads(grid_dims, group_dims);
+    compute_encoder.dispatch_threads(grid_dims, group_dims);
   } else {
     size_t dim0 = ndim > 0 ? shape[ndim - 1] : 1;
     size_t dim1 = ndim > 1 ? shape[ndim - 2] : 1;
     size_t rest = outputs[0].size() / (dim0 * dim1);
-    int work_per_thread = ndim > 3 ? WORK_PER_THREAD : 1;
+    int work_per_thread = ndim > 3 ? (large ? 4 : 2) : 1;
     dim0 = (dim0 + work_per_thread - 1) / work_per_thread;
     NS::UInteger thread_group_size = kernel->maxTotalThreadsPerThreadgroup();
     int pow2;
@@ -445,7 +503,7 @@ void Compiled::eval_gpu(
     }
     auto group_dims = get_block_dims(dim0, dim1, rest, pow2);
     MTL::Size grid_dims = MTL::Size(dim0, dim1, rest);
-    compute_encoder.dispatchThreads(grid_dims, group_dims);
+    compute_encoder.dispatch_threads(grid_dims, group_dims);
   }
 }
 

mlx/backend/metal/conv.cpp

@@ -44,27 +44,28 @@ void explicit_gemm_conv_ND_gpu(
   kname << "naive_unfold_nd_" << type_to_name(in_unfolded) << "_" << N;
   auto& compute_encoder = d.get_command_encoder(s.index);
   auto kernel = d.get_kernel(kname.str());
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
 
   compute_encoder.set_input_array(in, 0);
   compute_encoder.set_output_array(in_unfolded, 1);
 
-  compute_encoder->setBytes(&conv_params, sizeof(conv_params), 2);
+  compute_encoder.set_bytes(conv_params, 2);
 
   // Launch unfolding kernel
-  int tgp_x = std::min(conv_params.C, 64);
+  size_t tgp_x = std::min(conv_params.C, 64);
   tgp_x = 32 * ((tgp_x + 32 - 1) / 32);
-  int tgp_y = 256 / tgp_x;
+  size_t tgp_y = 256 / tgp_x;
 
-  MTL::Size group_dims = MTL::Size(tgp_x, tgp_y, 1);
   MTL::Size grid_dims = MTL::Size(
       conv_params.C, unfolded_shape[1] / conv_params.C, unfolded_shape[0]);
+  MTL::Size group_dims = MTL::Size(
+      std::min(tgp_x, grid_dims.width), std::min(tgp_y, grid_dims.height), 1);
 
-  compute_encoder.dispatchThreads(grid_dims, group_dims);
+  compute_encoder.dispatch_threads(grid_dims, group_dims);
 
   // Reshape weight
-  std::vector<int> wt_reshape{implicit_K, implicit_N};
+  Shape wt_reshape{implicit_K, implicit_N};
-  std::vector<size_t> wt_restride{1, static_cast<size_t>(implicit_K)};
+  Strides wt_restride{1, implicit_K};
   array wt_reshaped(wt_reshape, wt.dtype(), nullptr, {});
   auto wt_flags = wt.flags();
   wt_flags.row_contiguous = false;
@@ -122,33 +123,31 @@ void explicit_gemm_conv_group_ND_gpu(
         << N;
   auto& compute_encoder = d.get_command_encoder(s.index);
   auto kernel = d.get_kernel(kname.str());
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
 
   compute_encoder.set_input_array(in, 0);
   compute_encoder.set_output_array(in_unfolded, 1);
 
-  compute_encoder->setBytes(&conv_params, sizeof(conv_params), 2);
+  compute_encoder.set_bytes(conv_params, 2);
 
   // Launch unfolding kernel
-  int tgp_x = std::min(conv_params.C, 64);
+  size_t tgp_x = std::min(conv_params.C, 64);
   tgp_x = 32 * ((tgp_x + 32 - 1) / 32);
-  int tgp_y = 256 / tgp_x;
+  size_t tgp_y = 256 / tgp_x;
 
-  MTL::Size group_dims = MTL::Size(tgp_x, tgp_y, 1);
   MTL::Size grid_dims = MTL::Size(
       conv_params.C, unfolded_shape[1] / conv_params.C, unfolded_shape[0]);
+  MTL::Size group_dims = MTL::Size(
+      std::min(tgp_x, grid_dims.width), std::min(tgp_y, grid_dims.height), 1);
 
-  compute_encoder.dispatchThreads(grid_dims, group_dims);
+  compute_encoder.dispatch_threads(grid_dims, group_dims);
 
   // Transpose kernel weights so that we can slice them by contiguous chunks
   // of channel groups.
   array wt_view(
       {wt.shape(0), C_per_group, kernel_size}, wt.dtype(), nullptr, {});
   wt_view.copy_shared_buffer(
-      wt,
+      wt, {wt.strides(0), 1, C_per_group}, wt.flags(), wt.size());
-      {wt.strides(0), 1, static_cast<size_t>(C_per_group)},
-      wt.flags(),
-      wt.size());
 
   // Materialize
   auto wt_transpose = array(wt_view.shape(), wt_view.dtype(), nullptr, {});
@@ -237,7 +236,7 @@ void slow_conv_2D_gpu(
   // Encode and dispatch kernel
   auto& compute_encoder = d.get_command_encoder(s.index);
   auto kernel = d.get_kernel(kname.str());
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
 
   size_t n_pixels = conv_params.oS[0] * conv_params.oS[1];
 
@@ -252,8 +251,8 @@ void slow_conv_2D_gpu(
   compute_encoder.set_input_array(wt, 1);
   compute_encoder.set_output_array(out, 2);
 
-  compute_encoder->setBytes(&conv_params, sizeof(MLXConvParams<2>), 3);
+  compute_encoder.set_bytes(conv_params, 3);
-  compute_encoder.dispatchThreadgroups(grid_dims, group_dims);
+  compute_encoder.dispatch_threadgroups(grid_dims, group_dims);
 }
 
 void implicit_gemm_conv_2D_gpu(
@@ -352,7 +351,7 @@ void implicit_gemm_conv_2D_gpu(
       wn,
       n_channel_specialization,
       small_filter);
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
 
   // Deduce grid launch dimensions
   int tile = 1 << swizzle_log;
@@ -368,11 +367,11 @@ void implicit_gemm_conv_2D_gpu(
   compute_encoder.set_output_array(out, 2);
 
   // Encode params
-  compute_encoder->setBytes(&conv_params, sizeof(MLXConvParams<2>), 3);
+  compute_encoder.set_bytes(conv_params, 3);
-  compute_encoder->setBytes(&gemm_params, sizeof(ImplicitGemmConv2DParams), 4);
+  compute_encoder.set_bytes(gemm_params, 4);
 
   // Launch kernel
-  compute_encoder.dispatchThreadgroups(grid_dims, group_dims);
+  compute_encoder.dispatch_threadgroups(grid_dims, group_dims);
 }
 
 void implicit_gemm_conv_2D_general_gpu(
@@ -506,7 +505,7 @@ void implicit_gemm_conv_2D_general_gpu(
   auto& compute_encoder = d.get_command_encoder(s.index);
   auto kernel =
       get_steel_conv_general_kernel(d, kname.str(), out, bm, bn, bk, wm, wn);
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
 
   // Deduce grid launch dimensions
   int tile = 1 << swizzle_log;
@@ -523,17 +522,15 @@ void implicit_gemm_conv_2D_general_gpu(
   compute_encoder.set_output_array(out, 2);
 
   // Encode params
-  compute_encoder->setBytes(&conv_params, sizeof(MLXConvParams<2>), 3);
+  compute_encoder.set_bytes(conv_params, 3);
-  compute_encoder->setBytes(&gemm_params, sizeof(ImplicitGemmConv2DParams), 4);
+  compute_encoder.set_bytes(gemm_params, 4);
-  compute_encoder->setBytes(&jump_params, sizeof(Conv2DGeneralJumpParams), 5);
+  compute_encoder.set_bytes(jump_params, 5);
 
-  compute_encoder->setBytes(
+  compute_encoder.set_vector_bytes(base_h, 6);
-      base_h.data(), sizeof(Conv2DGeneralBaseInfo) * base_h.size(), 6);
+  compute_encoder.set_vector_bytes(base_w, 7);
-  compute_encoder->setBytes(
-      base_w.data(), sizeof(Conv2DGeneralBaseInfo) * base_w.size(), 7);
 
   // Launch kernel
-  compute_encoder.dispatchThreadgroups(grid_dims, group_dims);
+  compute_encoder.dispatch_threadgroups(grid_dims, group_dims);
 }
 
 void winograd_conv_2D_gpu(
@@ -622,18 +619,18 @@ void winograd_conv_2D_gpu(
           << bc;
     auto& compute_encoder = d.get_command_encoder(s.index);
     auto kernel = d.get_kernel(kname.str());
-    compute_encoder->setComputePipelineState(kernel);
+    compute_encoder.set_compute_pipeline_state(kernel);
 
     compute_encoder.set_input_array(wt, 0);
     compute_encoder.set_output_array(filt_wg, 1);
 
-    compute_encoder->setBytes(&C_c, sizeof(int), 2);
+    compute_encoder.set_bytes(C_c, 2);
-    compute_encoder->setBytes(&O_c, sizeof(int), 3);
+    compute_encoder.set_bytes(O_c, 3);
 
     MTL::Size group_dims = MTL::Size(32, bo, 1);
     MTL::Size grid_dims = MTL::Size(O_c / bo, 1, 1);
 
-    compute_encoder.dispatchThreadgroups(grid_dims, group_dims);
+    compute_encoder.dispatch_threadgroups(grid_dims, group_dims);
   }
 
   // Do input transform
@@ -650,18 +647,17 @@ void winograd_conv_2D_gpu(
           << bc;
     auto& compute_encoder = d.get_command_encoder(s.index);
     auto kernel = d.get_kernel(kname.str());
-    compute_encoder->setComputePipelineState(kernel);
+    compute_encoder.set_compute_pipeline_state(kernel);
 
     compute_encoder.set_input_array(in_padded, 0);
     compute_encoder.set_output_array(inp_wg, 1);
 
-    compute_encoder->setBytes(
+    compute_encoder.set_bytes(conv_params_updated, 2);
-        &conv_params_updated, sizeof(MLXConvParams<2>), 2);
 
     MTL::Size group_dims = MTL::Size(32, wn, wm);
     MTL::Size grid_dims = MTL::Size(N_tiles_w, N_tiles_h, N_tiles_n);
 
-    compute_encoder.dispatchThreadgroups(grid_dims, group_dims);
+    compute_encoder.dispatch_threadgroups(grid_dims, group_dims);
   }
 
   // Do batched gemm
@@ -698,18 +694,17 @@ void winograd_conv_2D_gpu(
           << bc;
     auto& compute_encoder = d.get_command_encoder(s.index);
     auto kernel = d.get_kernel(kname.str());
-    compute_encoder->setComputePipelineState(kernel);
+    compute_encoder.set_compute_pipeline_state(kernel);
 
     compute_encoder.set_input_array(out_wg, 0);
     compute_encoder.set_output_array(out, 1);
 
-    compute_encoder->setBytes(
+    compute_encoder.set_bytes(conv_params_updated, 2);
-        &conv_params_updated, sizeof(MLXConvParams<2>), 2);
 
     MTL::Size group_dims = MTL::Size(32, wn, wm);
     MTL::Size grid_dims = MTL::Size(N_tiles_w, N_tiles_h, N_tiles_n);
 
-    compute_encoder.dispatchThreadgroups(grid_dims, group_dims);
+    compute_encoder.dispatch_threadgroups(grid_dims, group_dims);
   }
 }
 

mlx/backend/metal/copy.cpp

@@ -43,13 +43,12 @@ void copy_gpu(const array& in, array& out, CopyType ctype) {
   copy_gpu(in, out, ctype, out.primitive().stream());
 }
 
-template <typename stride_t>
 void copy_gpu_inplace(
     const array& in,
     array& out,
-    const std::vector<int>& data_shape,
+    const Shape& data_shape,
-    const std::vector<stride_t>& strides_in_pre,
+    const Strides& strides_in_pre,
-    const std::vector<stride_t>& strides_out_pre,
+    const Strides& strides_out_pre,
     int64_t inp_offset,
     int64_t out_offset,
     CopyType ctype,
@@ -68,50 +67,52 @@ void copy_gpu_inplace(
               /* size_cap = */ INT32_MAX);
           return std::make_tuple(shape, strides[0], strides[1]);
         } else {
-          std::vector<stride_t> e;
+          Strides e{};
-          return std::make_tuple(std::vector<int>{}, e, e);
+          return std::make_tuple(Shape{}, e, e);
         }
       };
   auto [shape, strides_in_, strides_out_] = maybe_collapse();
   int ndim = shape.size();
-
+  bool large;
-  bool use_2d = out.data_size() > UINT32_MAX;
+  if (ctype == CopyType::General || ctype == CopyType::GeneralGeneral) {
+    // Allow for negative strides
+    large = in.data_size() > INT32_MAX || out.data_size() > INT32_MAX;
+  } else {
+    large = out.data_size() > UINT32_MAX;
+  }
   auto& d = metal::device(s.device);
   int work_per_thread = 1;
   std::string kernel_name;
-  {
+  switch (ctype) {
-    std::ostringstream kname;
+    case CopyType::Scalar:
-    switch (ctype) {
+      kernel_name = (large ? "s2" : "s");
-      case CopyType::Scalar:
+      break;
-        kname << (use_2d ? "s2" : "s");
+    case CopyType::Vector:
-        break;
+      kernel_name = (large ? "v2" : "v");
-      case CopyType::Vector:
+      break;
-        kname << (use_2d ? "v2" : "v");
+    case CopyType::General:
-        break;
+      kernel_name = "g";
-      case CopyType::General:
+      break;
-        kname << "g";
+    case CopyType::GeneralGeneral:
-        break;
+      kernel_name = "gg";
-      case CopyType::GeneralGeneral:
+      break;
-        kname << "gg";
-        break;
-    }
-    if (ctype == CopyType::General || ctype == CopyType::GeneralGeneral) {
-      if (shape.size() <= MAX_COPY_SPECIALIZED_DIMS) {
-        kname << shape.size();
-      } else {
-        work_per_thread = 4;
-        kname << "n4";
-      }
-    }
-    kname << "_copy";
-    kname << type_to_name(in) << type_to_name(out);
-    kernel_name = kname.str();
   }
-
+  if (ctype == CopyType::General || ctype == CopyType::GeneralGeneral) {
+    if (shape.size() <= MAX_COPY_SPECIALIZED_DIMS) {
+      kernel_name += std::to_string(shape.size());
+    } else {
+      work_per_thread = large ? 4 : 2;
+      concatenate(kernel_name, "n", std::to_string(work_per_thread));
+    }
+    if (large) {
+      kernel_name += "large";
+    }
+  }
+  concatenate(kernel_name, "_copy", type_to_name(in), type_to_name(out));
   auto kernel = get_copy_kernel(d, kernel_name, in, out);
 
   auto& compute_encoder = d.get_command_encoder(s.index);
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
   bool donate_in = in.data_shared_ptr() == nullptr;
 
   inp_offset *= size_of(in.dtype());
@@ -122,26 +123,26 @@ void copy_gpu_inplace(
 
   auto thread_group_size = kernel->maxTotalThreadsPerThreadgroup();
   if (ctype == CopyType::General || ctype == CopyType::GeneralGeneral) {
-    std::vector<int64_t> strides_in{strides_in_.begin(), strides_in_.end()};
+    Strides strides_in{strides_in_.begin(), strides_in_.end()};
-    std::vector<int64_t> strides_out{strides_out_.begin(), strides_out_.end()};
+    Strides strides_out{strides_out_.begin(), strides_out_.end()};
     if (ndim > 3) {
-      set_vector_bytes(compute_encoder, shape, ndim, 2);
+      compute_encoder.set_vector_bytes(shape, ndim, 2);
     }
-    set_vector_bytes(compute_encoder, strides_in, ndim, 3);
+    compute_encoder.set_vector_bytes(strides_in, ndim, 3);
     if (ctype == CopyType::GeneralGeneral) {
-      set_vector_bytes(compute_encoder, strides_out, ndim, 4);
+      compute_encoder.set_vector_bytes(strides_out, ndim, 4);
     }
 
-    int dim0 = ndim > 0 ? shape[ndim - 1] : 1;
+    size_t dim0 = ndim > 0 ? shape[ndim - 1] : 1;
-    int dim1 = ndim > 1 ? shape[ndim - 2] : 1;
+    size_t dim1 = ndim > 1 ? shape[ndim - 2] : 1;
 
     size_t data_size = 1;
     for (auto& s : shape)
       data_size *= s;
-    int rest = data_size / (dim0 * dim1);
+    size_t rest = data_size / (dim0 * dim1);
 
     if (ndim > MAX_COPY_SPECIALIZED_DIMS) {
-      compute_encoder->setBytes(&ndim, sizeof(int), 5);
+      compute_encoder.set_bytes(ndim, 5);
       dim0 = (dim0 + work_per_thread - 1) / work_per_thread;
     }
 
@@ -152,16 +153,16 @@ void copy_gpu_inplace(
 
     auto group_dims = get_block_dims(dim0, dim1, rest);
     MTL::Size grid_dims = MTL::Size(dim0, dim1, rest);
-    compute_encoder.dispatchThreads(grid_dims, group_dims);
+    compute_encoder.dispatch_threads(grid_dims, group_dims);
   } else {
     size_t nthreads = out.data_size();
     if (thread_group_size > nthreads) {
       thread_group_size = nthreads;
     }
     MTL::Size group_dims = MTL::Size(thread_group_size, 1, 1);
-    MTL::Size grid_dims = use_2d ? get_2d_grid_dims(out.shape(), out.strides())
+    MTL::Size grid_dims = large ? get_2d_grid_dims(out.shape(), out.strides())
-                                 : MTL::Size(nthreads, 1, 1);
+                                : MTL::Size(nthreads, 1, 1);
-    compute_encoder.dispatchThreads(grid_dims, group_dims);
+    compute_encoder.dispatch_threads(grid_dims, group_dims);
   }
 }
 
@@ -178,14 +179,13 @@ void copy_gpu_inplace(
 void copy_gpu_inplace(
     const array& in,
     array& out,
-    const std::vector<int64_t>& istride,
+    const Strides& istride,
     int64_t ioffset,
     CopyType ctype,
     const Stream& s) {
   assert(in.shape() == out.shape());
-  std::vector<int64_t> ostrides{out.strides().begin(), out.strides().end()};
   return copy_gpu_inplace(
-      in, out, in.shape(), istride, ostrides, ioffset, 0, ctype, s);
+      in, out, in.shape(), istride, out.strides(), ioffset, 0, ctype, s);
 }
 
 void fill_gpu(const array& val, array& out, const Stream& s) {
@@ -193,13 +193,13 @@ void fill_gpu(const array& val, array& out, const Stream& s) {
     return;
   }
   out.set_data(allocator::malloc_or_wait(out.nbytes()));
-  bool use_2d = out.data_size() > UINT32_MAX;
+  bool large = out.data_size() > UINT32_MAX;
   auto& d = metal::device(s.device);
-  std::string kernel_name = std::string(use_2d ? "s2" : "s") + "_copy" +
+  std::string kernel_name = std::string(large ? "s2" : "s") + "_copy" +
       type_to_name(val) + type_to_name(out);
   auto kernel = get_copy_kernel(d, kernel_name, val, out);
   auto& compute_encoder = d.get_command_encoder(s.index);
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
 
   compute_encoder.set_input_array(val, 0);
   compute_encoder.set_output_array(out, 1);
@@ -210,9 +210,9 @@ void fill_gpu(const array& val, array& out, const Stream& s) {
     thread_group_size = nthreads;
   }
   MTL::Size group_dims = MTL::Size(thread_group_size, 1, 1);
-  MTL::Size grid_dims = use_2d ? get_2d_grid_dims(out.shape(), out.strides())
+  MTL::Size grid_dims = large ? get_2d_grid_dims(out.shape(), out.strides())
-                               : MTL::Size(nthreads, 1, 1);
+                              : MTL::Size(nthreads, 1, 1);
-  compute_encoder.dispatchThreads(grid_dims, group_dims);
+  compute_encoder.dispatch_threads(grid_dims, group_dims);
 }
 
 } // namespace mlx::core

mlx/backend/metal/copy.h

@@ -8,13 +8,12 @@
 namespace mlx::core {
 
 // Generic copy inplace
-template <typename stride_t>
 void copy_gpu_inplace(
     const array& in,
     array& out,
-    const std::vector<int>& data_shape,
+    const Shape& data_shape,
-    const std::vector<stride_t>& i_strides,
+    const Strides& i_strides,
-    const std::vector<stride_t>& o_strides,
+    const Strides& o_strides,
     int64_t i_offset,
     int64_t o_offset,
     CopyType ctype,
@@ -32,7 +31,7 @@ void copy_gpu_inplace(
 void copy_gpu_inplace(
     const array& in,
     array& out,
-    const std::vector<int64_t>& istride,
+    const Strides& istride,
     int64_t ioffset,
     CopyType ctype,
     const Stream& s);

mlx/backend/metal/custom_kernel.cpp

@@ -43,7 +43,7 @@ void CustomKernel::eval_gpu(
       d.get_library(lib_name, [this] { return metal::utils() + source_; });
   auto kernel = d.get_kernel(name_, lib);
   auto& compute_encoder = d.get_command_encoder(s.index);
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
   int index = 0;
   for (int i = 0; i < checked_inputs.size(); i++) {
     const array& in = checked_inputs[i];
@@ -53,15 +53,15 @@ void CustomKernel::eval_gpu(
     if (in.ndim() > 0) {
       int ndim = in.ndim();
       if (shape_info.shape) {
-        set_vector_bytes(compute_encoder, in.shape(), ndim, index);
+        compute_encoder.set_vector_bytes(in.shape(), ndim, index);
         index++;
       }
       if (shape_info.strides) {
-        set_vector_bytes(compute_encoder, in.strides(), ndim, index);
+        compute_encoder.set_vector_bytes(in.strides(), ndim, index);
         index++;
       }
       if (shape_info.ndim) {
-        compute_encoder->setBytes(&ndim, sizeof(int), index);
+        compute_encoder.set_bytes(ndim, index);
         index++;
       }
     }
@@ -72,10 +72,11 @@ void CustomKernel::eval_gpu(
   }
 
   const auto [tx, ty, tz] = threadgroup_;
-  MTL::Size group_dims = MTL::Size(tx, ty, tz);
   const auto [gx, gy, gz] = grid_;
+  MTL::Size group_dims =
+      MTL::Size(std::min(tx, gx), std::min(ty, gy), std::min(tz, gz));
   MTL::Size grid_dims = MTL::Size(gx, gy, gz);
-  compute_encoder->dispatchThreads(grid_dims, group_dims);
+  compute_encoder.dispatch_threads(grid_dims, group_dims);
 
   d.add_temporaries(std::move(copies), s.index);
 }

mlx/backend/metal/device.cpp

@@ -23,14 +23,18 @@ constexpr int MAX_BUFFERS_PER_QUEUE = 12;
 
 constexpr const char* default_mtllib_path = METAL_PATH;
 
-constexpr auto get_metal_version() {
+auto get_metal_version() {
-#if (MLX_METAL_VERSION >= 320)
+  auto get_metal_version_ = []() {
-  return MTL::LanguageVersion3_2;
+    if (__builtin_available(macOS 15, iOS 18, tvOS 18, visionOS 2, *)) {
-#elif (MLX_METAL_VERSION >= 310)
+      return MTL::LanguageVersion3_2;
-  return MTL::LanguageVersion3_1;
+    } else if (__builtin_available(macOS 14, iOS 17, tvOS 17, visionOS 1, *)) {
-#else
+      return MTL::LanguageVersion3_1;
-  return MTL::LanguageVersion3_0;
+    } else {
-#endif
+      return MTL::LanguageVersion3_0;
+    }
+  };
+  static auto metal_version_ = get_metal_version_();
+  return metal_version_;
 }
 
 auto load_device() {
@@ -171,14 +175,14 @@ void CommandEncoder::maybeInsertBarrier() {
   next_outputs_.clear();
 }
 
-void CommandEncoder::dispatchThreadgroups(
+void CommandEncoder::dispatch_threadgroups(
     MTL::Size grid_dims,
     MTL::Size group_dims) {
   maybeInsertBarrier();
   enc_->dispatchThreadgroups(grid_dims, group_dims);
 }
 
-void CommandEncoder::dispatchThreads(
+void CommandEncoder::dispatch_threads(
     MTL::Size grid_dims,
     MTL::Size group_dims) {
   maybeInsertBarrier();
@@ -276,7 +280,7 @@ void Device::end_encoding(int index) {
     // - Update the map of outputs to include this command encoder's outputs.
     // - Always signal this command encoders fence.
     // - Add a completion handler for this command encoder that removes outputs
-    //   from the map to limit the growth of the map and avoid unecessary waits
+    //   from the map to limit the growth of the map and avoid unnecessary waits
     // - Temporaries are a special case as they do not cross command encoder
     //   boundaries. These can be removed early from the encoders inputs and
     //   outputs since they don't need synchronization.
@@ -298,7 +302,7 @@ void Device::end_encoding(int index) {
         if (auto it = stream.outputs.find(in); it != stream.outputs.end()) {
           // If we've already waited on a fence, don't wait on it again.
           if (waiting_on.find(it->second) == waiting_on.end()) {
-            enc->waitForFence(it->second->fence);
+            enc.wait_for_fence(it->second->fence);
             waiting_on.insert(it->second);
           }
         }
@@ -307,7 +311,7 @@ void Device::end_encoding(int index) {
         stream.outputs[out] = stream.fence;
       }
     }
-    enc->updateFence(stream.fence->fence);
+    enc.update_fence(stream.fence->fence);
     stream.buffer->addCompletedHandler(
         [&stream,
          waiting_on = std::move(waiting_on),
@@ -641,21 +645,27 @@ void new_stream(Stream stream) {
 
 std::unordered_map<std::string, std::variant<std::string, size_t>>
 device_info() {
-  auto raw_device = device(default_device()).mtl_device();
+  auto init_device_info = []()
-  auto arch = std::string(raw_device->architecture()->name()->utf8String());
+      -> std::unordered_map<std::string, std::variant<std::string, size_t>> {
+    auto pool = new_scoped_memory_pool();
+    auto raw_device = device(default_device()).mtl_device();
+    auto arch = std::string(raw_device->architecture()->name()->utf8String());
 
-  int mib[] = {CTL_HW, HW_MEMSIZE};
+    int mib[] = {CTL_HW, HW_MEMSIZE};
-  size_t memsize = 0;
+    size_t memsize = 0;
-  size_t length = sizeof(memsize);
+    size_t length = sizeof(memsize);
 
-  sysctl(mib, 2, &memsize, &length, NULL, 0);
+    sysctl(mib, 2, &memsize, &length, NULL, 0);
 
-  return {
+    return {
-      {"architecture", arch},
+        {"architecture", arch},
-      {"max_buffer_length", raw_device->maxBufferLength()},
+        {"max_buffer_length", raw_device->maxBufferLength()},
-      {"max_recommended_working_set_size",
+        {"max_recommended_working_set_size",
-       raw_device->recommendedMaxWorkingSetSize()},
+         raw_device->recommendedMaxWorkingSetSize()},
-      {"memory_size", memsize}};
+        {"memory_size", memsize}};
+  };
+  static auto device_info_ = init_device_info();
+  return device_info_;
 }
 
 } // namespace mlx::core::metal

mlx/backend/metal/device.h

@@ -58,16 +58,43 @@ struct CommandEncoder {
     CommandEncoder& enc;
   };
 
-  MTL::ComputeCommandEncoder* operator->() {
-    return enc_;
-  }
-
   void set_input_array(const array& a, int idx, int64_t offset = 0);
   void set_output_array(array& a, int idx, int64_t offset = 0);
-  void dispatchThreadgroups(MTL::Size grid_dims, MTL::Size group_dims);
+  void dispatch_threadgroups(MTL::Size grid_dims, MTL::Size group_dims);
-  void dispatchThreads(MTL::Size grid_dims, MTL::Size group_dims);
+  void dispatch_threads(MTL::Size grid_dims, MTL::Size group_dims);
   void maybeInsertBarrier();
 
+  void set_compute_pipeline_state(MTL::ComputePipelineState* kernel) {
+    enc_->setComputePipelineState(kernel);
+  }
+
+  void wait_for_fence(MTL::Fence* fence) {
+    enc_->waitForFence(fence);
+  }
+
+  void update_fence(MTL::Fence* fence) {
+    enc_->updateFence(fence);
+  }
+
+  template <typename T>
+  void set_vector_bytes(const std::vector<T>& vec, size_t nelems, int idx) {
+    enc_->setBytes(vec.data(), nelems * sizeof(T), idx);
+  }
+  template <typename T>
+  void set_vector_bytes(const std::vector<T>& vec, int idx) {
+    return set_vector_bytes(vec, vec.size(), idx);
+  }
+
+  template <typename T>
+  void set_bytes(const T* v, int n, int idx) {
+    return enc_->setBytes(v, n * sizeof(T), idx);
+  }
+
+  template <typename T>
+  void set_bytes(const T& v, int idx) {
+    return enc_->setBytes(&v, sizeof(T), idx);
+  }
+
   ConcurrentContext start_concurrent() {
     return ConcurrentContext(*this);
   }

mlx/backend/metal/fft.cpp

@@ -363,7 +363,7 @@ void multi_upload_bluestein_fft(
   auto [w_k, w_q] = compute_bluestein_constants(n, plan.bluestein_n);
 
   // Broadcast w_q and w_k to the batch size
-  std::vector<size_t> b_strides(in.ndim(), 0);
+  Strides b_strides(in.ndim(), 0);
   b_strides[axis] = 1;
   array w_k_broadcast({}, complex64, nullptr, {});
   array w_q_broadcast({}, complex64, nullptr, {});
@@ -386,8 +386,8 @@ void multi_upload_bluestein_fft(
     copies.push_back(slice_temp);
     copies.push_back(conj_temp);
 
-    std::vector<int> rstarts(in.ndim(), 0);
+    Shape rstarts(in.ndim(), 0);
-    std::vector<int> rstrides(in.ndim(), 1);
+    Shape rstrides(in.ndim(), 1);
     rstarts[axis] = in.shape(axis) - back_offset;
     rstrides[axis] = -1;
     unary_op_gpu({in}, conj_temp, "Conjugate", s);
@@ -431,19 +431,19 @@ void multi_upload_bluestein_fft(
       s);
 
   int offset = plan.bluestein_n - (2 * n - 1);
-  std::vector<int> starts(in.ndim(), 0);
+  Shape starts(in.ndim(), 0);
-  std::vector<int> strides(in.ndim(), 1);
+  Shape strides(in.ndim(), 1);
   starts[axis] = plan.bluestein_n - offset - n;
   slice_gpu(pad_temp1, temp, starts, strides, s);
 
   binary_op_gpu_inplace({temp, w_k_broadcast}, temp1, "Multiply", s);
 
   if (real && !inverse) {
-    std::vector<int> rstarts(in.ndim(), 0);
+    Shape rstarts(in.ndim(), 0);
-    std::vector<int> rstrides(in.ndim(), 1);
+    Shape rstrides(in.ndim(), 1);
     slice_gpu(temp1, out, rstarts, strides, s);
   } else if (real && inverse) {
-    std::vector<size_t> b_strides(in.ndim(), 0);
+    Strides b_strides(in.ndim(), 0);
     auto inv_n = array({1.0f / n}, {1}, float32);
     array temp_float(out.shape(), out.dtype(), nullptr, {});
     copies.push_back(temp_float);
@@ -531,8 +531,8 @@ void fft_op(
       return x;
     } else {
       array x_copy(x.shape(), x.dtype(), nullptr, {});
-      std::vector<size_t> strides;
+      Strides strides;
-      size_t cur_stride = x.shape(axis);
+      int64_t cur_stride = x.shape(axis);
       for (int a = 0; a < x.ndim(); a++) {
         if (a == axis) {
           strides.push_back(1);
@@ -699,7 +699,7 @@ void fft_op(
     auto kernel =
         get_fft_kernel(d, base_name, hash_name, func_consts, template_def);
 
-    compute_encoder->setComputePipelineState(kernel);
+    compute_encoder.set_compute_pipeline_state(kernel);
     compute_encoder.set_input_array(in_contiguous, 0);
     compute_encoder.set_output_array(out, 1);
 
@@ -711,9 +711,9 @@ void fft_op(
 
       compute_encoder.set_input_array(w_q, 2); // w_q
       compute_encoder.set_input_array(w_k, 3); // w_k
-      compute_encoder->setBytes(&n, sizeof(int), 4);
+      compute_encoder.set_bytes(n, 4);
-      compute_encoder->setBytes(&plan.bluestein_n, sizeof(int), 5);
+      compute_encoder.set_bytes(plan.bluestein_n, 5);
-      compute_encoder->setBytes(&total_batch_size, sizeof(int), 6);
+      compute_encoder.set_bytes(total_batch_size, 6);
     } else if (plan.rader_n > 1) {
       auto [b_q, g_q, g_minus_q] = compute_raders_constants(plan.rader_n, s);
       copies.push_back(b_q);
@@ -723,22 +723,22 @@ void fft_op(
       compute_encoder.set_input_array(b_q, 2);
       compute_encoder.set_input_array(g_q, 3);
       compute_encoder.set_input_array(g_minus_q, 4);
-      compute_encoder->setBytes(&n, sizeof(int), 5);
+      compute_encoder.set_bytes(n, 5);
-      compute_encoder->setBytes(&total_batch_size, sizeof(int), 6);
+      compute_encoder.set_bytes(total_batch_size, 6);
-      compute_encoder->setBytes(&plan.rader_n, sizeof(int), 7);
+      compute_encoder.set_bytes(plan.rader_n, 7);
     } else if (four_step_params.required) {
-      compute_encoder->setBytes(&four_step_params.n1, sizeof(int), 2);
+      compute_encoder.set_bytes(four_step_params.n1, 2);
-      compute_encoder->setBytes(&four_step_params.n2, sizeof(int), 3);
+      compute_encoder.set_bytes(four_step_params.n2, 3);
-      compute_encoder->setBytes(&total_batch_size, sizeof(int), 4);
+      compute_encoder.set_bytes(total_batch_size, 4);
     } else {
-      compute_encoder->setBytes(&n, sizeof(int), 2);
+      compute_encoder.set_bytes(n, 2);
-      compute_encoder->setBytes(&total_batch_size, sizeof(int), 3);
+      compute_encoder.set_bytes(total_batch_size, 3);
     }
 
     auto group_dims = MTL::Size(1, threadgroup_batch_size, threads_per_fft);
     auto grid_dims =
         MTL::Size(batch_size, threadgroup_batch_size, threads_per_fft);
-    compute_encoder->dispatchThreads(grid_dims, group_dims);
+    compute_encoder.dispatch_threads(grid_dims, group_dims);
   }
 
   d.add_temporaries(std::move(copies), s.index);
@@ -777,7 +777,7 @@ void nd_fft_op(
     // Mirror np.fft.(i)rfftn and perform a real transform
     // only on the final axis.
     bool step_real = (real && index == axes.size() - 1);
-    int step_shape = inverse ? out.shape(axis) : in.shape(axis);
+    auto step_shape = inverse ? out.shape(axis) : in.shape(axis);
     const array& in_arr = i == axes.size() - 1 ? in : temp_arrs[1 - i % 2];
     array& out_arr = i == 0 ? out : temp_arrs[i % 2];
     fft_op(in_arr, out_arr, axis, inverse, step_real, inplace, s);

mlx/backend/metal/hadamard.cpp

@@ -137,14 +137,14 @@ void Hadamard::eval_gpu(const std::vector<array>& inputs, array& out) {
     auto kernel = d.get_kernel(kernel_name, lib);
     assert(threads_per <= kernel->maxTotalThreadsPerThreadgroup());
 
-    compute_encoder->setComputePipelineState(kernel);
+    compute_encoder.set_compute_pipeline_state(kernel);
     compute_encoder.set_input_array(in, 0);
     compute_encoder.set_output_array(out, 1);
-    compute_encoder->setBytes(&scale, sizeof(float), 2);
+    compute_encoder.set_bytes(scale, 2);
 
     MTL::Size group_dims = MTL::Size(1, threads_per, 1);
     MTL::Size grid_dims = MTL::Size(batch_size, threads_per, 1);
-    compute_encoder->dispatchThreads(grid_dims, group_dims);
+    compute_encoder.dispatch_threads(grid_dims, group_dims);
   };
 
   if (m > 1) {

mlx/backend/metal/indexing.cpp

@@ -53,27 +53,31 @@ void Gather::eval_gpu(const std::vector<array>& inputs, array& out) {
   int idx_ndim = nidx ? inputs[1].ndim() : 0;
   size_t ndim = src.ndim();
 
-  std::string lib_name;
+  bool large_index = nidx && inputs[1].size() > UINT32_MAX;
-  std::string kernel_name;
+  bool large_src = src.size() > UINT32_MAX;
+  bool large_out = out.size() > UINT32_MAX;
+  bool large = large_index || large_src || large_out;
+
   std::string idx_type_name = nidx ? type_to_name(inputs[1]) : "";
-  {
+  std::string kernel_name = fmt::format(
-    std::ostringstream kname;
+      "gather{0}{1}_{2}_{3}_{4}",
-    kname << "gather" << type_to_name(out) << idx_type_name << "_" << nidx
+      type_to_name(out),
-          << "_" << idx_ndim;
+      idx_type_name,
-    lib_name = kname.str();
+      nidx,
-    kernel_name = lib_name;
+      idx_ndim,
-  }
+      large ? "int64_t" : "uint");
+  std::string lib_name = kernel_name;
 
   auto lib = d.get_library(lib_name, [&]() {
-    std::ostringstream kernel_source;
+    std::string kernel_source = metal::utils();
-    kernel_source << metal::utils() << metal::gather();
+    kernel_source += metal::gather();
     std::string out_type_str = get_type_string(out.dtype());
     std::string idx_type_str =
         nidx ? get_type_string(inputs[1].dtype()) : "bool";
     auto [idx_args, idx_arr] = make_index_args(idx_type_str, nidx);
 
     // Index dimension specializations
-    kernel_source << fmt::format(
+    kernel_source += fmt::format(
         gather_kernels,
         type_to_name(out) + idx_type_name,
         out_type_str,
@@ -81,13 +85,14 @@ void Gather::eval_gpu(const std::vector<array>& inputs, array& out) {
         nidx,
         idx_args,
         idx_arr,
-        idx_ndim);
+        idx_ndim,
-    return kernel_source.str();
+        large ? "int64_t" : "uint");
+    return kernel_source;
   });
 
   auto& compute_encoder = d.get_command_encoder(s.index);
   auto kernel = d.get_kernel(kernel_name, lib);
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
 
   size_t slice_size = 1;
   for (auto s : slice_sizes_) {
@@ -131,20 +136,20 @@ void Gather::eval_gpu(const std::vector<array>& inputs, array& out) {
   compute_encoder.set_output_array(out, 1);
 
   // Set source info
-  set_vector_bytes(compute_encoder, src.shape(), 2);
+  compute_encoder.set_vector_bytes(src.shape(), 2);
-  set_vector_bytes(compute_encoder, src.strides(), 3);
+  compute_encoder.set_vector_bytes(src.strides(), 3);
-  compute_encoder->setBytes(&ndim, sizeof(size_t), 4);
+  compute_encoder.set_bytes(ndim, 4);
-  set_vector_bytes(compute_encoder, slice_sizes_, 5);
+  compute_encoder.set_vector_bytes(slice_sizes_, 5);
-  set_vector_bytes(compute_encoder, axes_, 6);
+  compute_encoder.set_vector_bytes(axes_, 6);
 
   // Set index info
   //
   // We don't need to check for empty idx_shapes because gather has a
   // idx_ndim == 0 specialization
-  set_vector_bytes(compute_encoder, idx_shapes, 7);
+  compute_encoder.set_vector_bytes(idx_shapes, 7);
-  set_vector_bytes(compute_encoder, idx_strides, 8);
+  compute_encoder.set_vector_bytes(idx_strides, 8);
-  set_vector_bytes(compute_encoder, idx_contigs, 9);
+  compute_encoder.set_vector_bytes(idx_contigs, 9);
-  compute_encoder->setBytes(&idx_ndim, sizeof(int), 10);
+  compute_encoder.set_bytes(idx_ndim, 10);
 
   // Set index buffers
   for (int i = 0; i < nidx; ++i) {
@@ -152,7 +157,7 @@ void Gather::eval_gpu(const std::vector<array>& inputs, array& out) {
   }
 
   // Launch grid
-  compute_encoder.dispatchThreads(grid_dims, group_dims);
+  compute_encoder.dispatch_threads(grid_dims, group_dims);
 }
 
 void Scatter::eval_gpu(const std::vector<array>& inputs, array& out) {
@@ -209,8 +214,6 @@ void Scatter::eval_gpu(const std::vector<array>& inputs, array& out) {
     nwork = 32;
   }
 
-  std::string lib_name;
-  std::string kernel_name;
   std::string idx_type_name = nidx ? type_to_name(inputs[1]) : "";
   std::string op_name;
   switch (reduce_type_) {
@@ -231,18 +234,24 @@ void Scatter::eval_gpu(const std::vector<array>& inputs, array& out) {
       break;
   }
   auto upd_contig = upd.flags().row_contiguous;
-  {
+  bool large_out = out.size() > UINT32_MAX;
-    std::ostringstream kname;
+  bool large_idx = nidx && (inputs[1].size() > UINT32_MAX);
-    kname << "scatter" << type_to_name(out) << idx_type_name;
+  bool large_upd = upd.size() > UINT32_MAX;
-    kname << "_" << op_name << "_" << nidx << "_"
+  bool large = large_out || large_idx || large_upd;
-          << (upd_contig ? "updc_true" : "updc_false") << "_nwork" << nwork;
+  std::string kernel_name = fmt::format(
-    lib_name = kname.str();
+      "scatter{0}{1}_{2}_{3}_{4}_nwork{5}_{6}",
-    kernel_name = kname.str();
+      type_to_name(out),
-  }
+      idx_type_name,
+      op_name,
+      nidx,
+      upd_contig ? "updc_true" : "updc_false",
+      nwork,
+      large ? "int64_t" : "uint");
+  std::string lib_name = kernel_name;
+
   auto lib = d.get_library(lib_name, [&]() {
-    std::ostringstream kernel_source;
+    std::string kernel_source = metal::utils();
-    kernel_source << metal::utils() << metal::reduce_utils()
+    concatenate(kernel_source, metal::reduce_utils(), metal::scatter());
-                  << metal::scatter();
 
     std::string out_type_str = get_type_string(out.dtype());
     std::string idx_type_str =
@@ -270,7 +279,7 @@ void Scatter::eval_gpu(const std::vector<array>& inputs, array& out) {
     }
     auto [idx_args, idx_arr] = make_index_args(idx_type_str, nidx);
 
-    kernel_source << fmt::format(
+    kernel_source += fmt::format(
         scatter_kernels,
         type_to_name(out) + idx_type_name + "_" + op_name,
         out_type_str,
@@ -280,8 +289,9 @@ void Scatter::eval_gpu(const std::vector<array>& inputs, array& out) {
         idx_args,
         idx_arr,
         upd_contig,
-        nwork);
+        nwork,
-    return kernel_source.str();
+        large ? "int64_t" : "uint");
+    return kernel_source;
   });
 
   auto& compute_encoder = d.get_command_encoder(s.index);
@@ -289,7 +299,7 @@ void Scatter::eval_gpu(const std::vector<array>& inputs, array& out) {
 
   size_t nthreads = upd.size();
 
-  compute_encoder->setComputePipelineState(kernel);
+  compute_encoder.set_compute_pipeline_state(kernel);
 
   // Set all the buffers
   compute_encoder.set_input_array(upd, 1);
@@ -302,8 +312,8 @@ void Scatter::eval_gpu(const std::vector<array>& inputs, array& out) {
     upd_size *= upd.shape(i);
   }
   // Collect all idx shapes and strides into one place
-  std::vector<int> idx_shapes;
+  Shape idx_shapes;
-  std::vector<size_t> idx_strides;
+  Strides idx_strides;
   // To access .data() use char instead of bool
   // bool is 1 byte in Metal so this is safe
   std::vector<char> idx_contigs;
@@ -322,30 +332,30 @@ void Scatter::eval_gpu(const std::vector<array>& inputs, array& out) {
   if (upd_ndim == 0) {
     // Need placeholders so Metal doesn't compalain
     int shape_ = 0;
-    size_t stride_ = 0;
+    int64_t stride_ = 0;
-    compute_encoder->setBytes(&shape_, sizeof(int), 3);
+    compute_encoder.set_bytes(shape_, 3);
-    compute_encoder->setBytes(&stride_, sizeof(size_t), 4);
+    compute_encoder.set_bytes(stride_, 4);
   } else {
-    set_vector_bytes(compute_encoder, upd.shape(), 3);
+    compute_encoder.set_vector_bytes(upd.shape(), 3);
-    set_vector_bytes(compute_encoder, upd.strides(), 4);
+    compute_encoder.set_vector_bytes(upd.strides(), 4);
   }
-  compute_encoder->setBytes(&upd_ndim, sizeof(size_t), 5);
+  compute_encoder.set_bytes(upd_ndim, 5);
-  compute_encoder->setBytes(&upd_size, sizeof(size_t), 6);
+  compute_encoder.set_bytes(upd_size, 6);
 
   // Set output info
   size_t out_ndim = out.ndim();
   if (out_ndim == 0) {
     // Need placeholders so Metal doesn't compalain
     int shape_ = 0;
-    size_t stride_ = 0;
+    int64_t stride_ = 0;
-    compute_encoder->setBytes(&shape_, sizeof(int), 7);
+    compute_encoder.set_bytes(shape_, 7);
-    compute_encoder->setBytes(&stride_, sizeof(size_t), 8);
+    compute_encoder.set_bytes(stride_, 8);
   } else {
-    set_vector_bytes(compute_encoder, out.shape(), 7);
+    compute_encoder.set_vector_bytes(out.shape(), 7);
-    set_vector_bytes(compute_encoder, out.strides(), 8);
+    compute_encoder.set_vector_bytes(out.strides(), 8);
   }
-  compute_encoder->setBytes(&out_ndim, sizeof(size_t), 9);
+  compute_encoder.set_bytes(out_ndim, 9);
-  compute_encoder->setBytes(axes_.data(), axes_.size() * sizeof(int), 10);
+  compute_encoder.set_vector_bytes(axes_, 10);
 
   // Set index info
   if (idx_ndim == 0) {
@@ -355,11 +365,11 @@ void Scatter::eval_gpu(const std::vector<array>& inputs, array& out) {
     idx_strides.push_back(0);
     idx_contigs.push_back(false);
   }
-  set_vector_bytes(compute_encoder, idx_shapes, 11);
+  compute_encoder.set_vector_bytes(idx_shapes, 11);
-  set_vector_bytes(compute_encoder, idx_strides, 12);
+  compute_encoder.set_vector_bytes(idx_strides, 12);
-  set_vector_bytes(compute_encoder, idx_contigs, 13);
+  compute_encoder.set_vector_bytes(idx_contigs, 13);
-  compute_encoder->setBytes(&idx_ndim, sizeof(int), 14);
+  compute_encoder.set_bytes(idx_ndim, 14);
-  compute_encoder->setBytes(&idx_size, sizeof(size_t), 15);
+  compute_encoder.set_bytes(idx_size, 15);
 
   // Set index buffers
   for (int i = 0; i < nidx; ++i) {
@@ -375,7 +385,7 @@ void Scatter::eval_gpu(const std::vector<array>& inputs, array& out) {
     throw std::runtime_error("[Scatter::eval_gpu] Invalid number of threads");
   }
   MTL::Size group_dims = get_block_dims(upd_size, grid_y, 1);
-  compute_encoder.dispatchThreads(grid_dims, group_dims);
+  compute_encoder.dispatch_threads(grid_dims, group_dims);
 }
 
 } // namespace mlx::core

mlx/backend/metal/jit/gemv_masked.h

@@ -11,13 +11,13 @@ gemv_{trans}masked<{itype}, {outm_t}, {opm_t}, {bm}, {bn}, {sm}, {sn}, {tm}, {tn
     const constant int& marix_ld [[buffer(6)]],
     const constant int& batch_ndim [[buffer(9)]],
     const constant int* batch_shape [[buffer(10)]],
-    const constant size_t* vector_batch_stride [[buffer(11)]],
+    const constant int64_t* vector_batch_stride [[buffer(11)]],
-    const constant size_t* matrix_batch_stride [[buffer(12)]],
+    const constant int64_t* matrix_batch_stride [[buffer(12)]],
     const device {outm_t}* out_mask [[buffer(20)]],
     const device {opm_t}* mat_mask [[buffer(21)]],
     const device {opm_t}* vec_mask [[buffer(22)]],
     const constant int* mask_strides [[buffer(23)]],
-    const constant size_t* mask_batch_strides [[buffer(24)]],
+    const constant int64_t* mask_batch_strides [[buffer(24)]],
     uint3 tid [[threadgroup_position_in_grid]],
     uint3 lid [[thread_position_in_threadgroup]],
     uint simd_gid [[simdgroup_index_in_threadgroup]],

mlx/backend/metal/jit/indexing.h

@@ -1,16 +1,16 @@
 // Copyright © 2023-2024 Apple Inc.
 
 constexpr std::string_view gather_kernels = R"(
-[[kernel]] void gather{0}_{3}_{6}(
+[[kernel]] void gather{0}_{3}_{6}_{7}(
     const device {1}* src [[buffer(0)]],
     device {1}* out [[buffer(1)]],
     const constant int* src_shape [[buffer(2)]],
-    const constant size_t* src_strides [[buffer(3)]],
+    const constant int64_t* src_strides [[buffer(3)]],
     const constant size_t& src_ndim [[buffer(4)]],
     const constant int* slice_sizes [[buffer(5)]],
     const constant int* axes [[buffer(6)]],
     const constant int* idx_shapes [[buffer(7)]],
-    const constant size_t* idx_strides [[buffer(8)]],
+    const constant int64_t* idx_strides [[buffer(8)]],
     const constant bool* idx_contigs [[buffer(9)]],
     const constant int& idx_ndim [[buffer(10)]],
     {4}
@@ -19,7 +19,7 @@ constexpr std::string_view gather_kernels = R"(
   Indices<{2}, {3}> idxs{{
     {{ {5} }}, idx_shapes, idx_strides, idx_contigs, idx_ndim}};
 
-  return gather_impl<{1}, {2}, {3}, {6}>(
+  return gather_impl<{1}, {2}, {3}, {6}, {7}>(
       src,
       out,
       src_shape,
@@ -34,19 +34,19 @@ constexpr std::string_view gather_kernels = R"(
 )";
 
 constexpr std::string_view scatter_kernels = R"(
-[[kernel]] void scatter{0}_{4}_updc_{7}_nwork{8}(
+[[kernel]] void scatter{0}_{4}_updc_{7}_nwork{8}_{9}(
     const device {1}* updates [[buffer(1)]],
     device mlx_atomic<{1}>* out [[buffer(2)]],
     const constant int* upd_shape [[buffer(3)]],
-    const constant size_t* upd_strides [[buffer(4)]],
+    const constant int64_t* upd_strides [[buffer(4)]],
     const constant size_t& upd_ndim [[buffer(5)]],
     const constant size_t& upd_size [[buffer(6)]],
     const constant int* out_shape [[buffer(7)]],
-    const constant size_t* out_strides [[buffer(8)]],
+    const constant int64_t* out_strides [[buffer(8)]],
     const constant size_t& out_ndim [[buffer(9)]],
     const constant int* axes [[buffer(10)]],
     const constant int* idx_shapes [[buffer(11)]],
-    const constant size_t* idx_strides [[buffer(12)]],
+    const constant int64_t* idx_strides [[buffer(12)]],
     const constant bool* idx_contigs [[buffer(13)]],
     const constant int& idx_ndim [[buffer(14)]],
     const constant size_t& idx_size [[buffer(15)]],
@@ -54,7 +54,7 @@ constexpr std::string_view scatter_kernels = R"(
     uint2 gid [[thread_position_in_grid]]) {{
   Indices<{2}, {4}> idxs{{ {{ {6} }}, idx_shapes, idx_strides, idx_contigs, idx_ndim}};
 
-  return scatter_impl<{1}, {2}, {3}, {4}, {7}, {8}>(
+  return scatter_impl<{1}, {2}, {3}, {4}, {7}, {8}, {9}>(
       updates,
       out,
       upd_shape,

mlx/backend/metal/jit/steel_gemm.h

@@ -10,12 +10,12 @@ template [[host_name("{name}")]]
     const constant GEMMParams* params [[buffer(4)]],
     const constant GEMMAddMMParams* addmm_params [[buffer(5), function_constant(use_out_source)]],
     const constant int* batch_shape [[buffer(6)]],
-    const constant size_t* batch_strides [[buffer(7)]],
+    const constant int64_t* batch_strides [[buffer(7)]],
     const constant uint32_t* lhs_indices [[buffer(10), function_constant(do_gather)]],
     const constant uint32_t* rhs_indices [[buffer(11), function_constant(do_gather)]],
     const constant uint32_t* C_indices [[buffer(12), function_constant(gather_bias)]],
     const constant int* operand_shape [[buffer(13), function_constant(do_gather)]],
-    const constant size_t* operand_strides [[buffer(14), function_constant(do_gather)]],
+    const constant int64_t* operand_strides [[buffer(14), function_constant(do_gather)]],
     const constant packed_int3& operand_batch_ndim [[buffer(15), function_constant(do_gather)]],
     uint simd_lane_id [[thread_index_in_simdgroup]],
     uint simd_group_id [[simdgroup_index_in_threadgroup]],
@@ -43,7 +43,7 @@ block_masked_gemm<
     device {itype}* D [[buffer(3)]],
     const constant GEMMParams* params [[buffer(4)]],
     const constant int* batch_shape [[buffer(6)]],
-    const constant size_t* batch_strides [[buffer(7)]],
+    const constant int64_t* batch_strides [[buffer(7)]],
     const device {outmasktype}* out_mask [[buffer(10)]],
     const device {opmasktype}* lhs_mask [[buffer(11)]],
     const device {opmasktype}* rhs_mask [[buffer(12)]],

mlx/backend/metal/jit_kernels.cpp

@@ -1,5 +1,4 @@
 // Copyright © 2024 Apple Inc.
-
 #include "mlx/backend/common/compiled.h"
 #include "mlx/backend/metal/jit/arange.h"
 #include "mlx/backend/metal/jit/gemv_masked.h"
@@ -46,25 +45,27 @@ MTL::ComputePipelineState* get_unary_kernel(
   auto lib = d.get_library(lib_name, [&]() {
     auto in_t = get_type_string(in_type);
     auto out_t = get_type_string(out_type);
-    std::ostringstream kernel_source;
+    std::string kernel_source = metal::utils();
-    kernel_source << metal::utils() << metal::unary_ops() << metal::unary();
+    concatenate(kernel_source, metal::unary_ops(), metal::unary());
-    kernel_source << get_template_definition(
+    kernel_source +=
-        "v_" + lib_name, "unary_v", in_t, out_t, op);
+        get_template_definition("v_" + lib_name, "unary_v", in_t, out_t, op);
-    kernel_source << get_template_definition(
+    kernel_source +=
-        "v2_" + lib_name, "unary_v2", in_t, out_t, op);
+        get_template_definition("v2_" + lib_name, "unary_v2", in_t, out_t, op);
-    kernel_source << get_template_definition(
+    kernel_source += get_template_definition(
-        "gn4_" + lib_name, "unary_g", in_t, out_t, op, 4);
+        "gn1_" + lib_name, "unary_g", in_t, out_t, op, 1, "uint");
-    return kernel_source.str();
+    kernel_source += get_template_definition(
+        "gn4large_" + lib_name, "unary_g", in_t, out_t, op, 4);
+    return kernel_source;
   });
   return d.get_kernel(kernel_name, lib);
 }
 
-void add_binary_kernels(
+void append_binary_kernels(
     const std::string lib_name,
     Dtype in_type,
     Dtype out_type,
     const std::string op,
-    std::ostringstream& kernel_source) {
+    std::string& kernel_source) {
   const std::array<std::pair<std::string, std::string>, 10> kernel_types = {{
       {"ss", "binary_ss"},
       {"vs", "binary_vs"},
@@ -74,26 +75,24 @@ void add_binary_kernels(
       {"sv2", "binary_sv2"},
       {"vv2", "binary_vv2"},
       {"g1", "binary_g_nd1"},
-      {"g2", "binary_g_nd2"},
+      {"g2large", "binary_g_nd2"},
-      {"g3", "binary_g_nd3"},
+      {"g3large", "binary_g_nd3"},
   }};
+  auto in_t = get_type_string(in_type);
+  auto out_t = get_type_string(out_type);
+
   for (auto& [name, func] : kernel_types) {
-    std::string template_def;
+    kernel_source +=
-    template_def = get_template_definition(
+        get_template_definition(name + "_" + lib_name, func, in_t, out_t, op);
-        name + "_" + lib_name,
-        func,
-        get_type_string(in_type),
-        get_type_string(out_type),
-        op);
-    kernel_source << template_def;
   }
-  kernel_source << get_template_definition(
+  kernel_source += get_template_definition(
-      "gn4_" + lib_name,
+      "g2_" + lib_name, "binary_g_nd2", in_t, out_t, op, "uint");
-      "binary_g",
+  kernel_source += get_template_definition(
-      get_type_string(in_type),
+      "g3_" + lib_name, "binary_g_nd3", in_t, out_t, op, "uint");
-      get_type_string(out_type),
+  kernel_source += get_template_definition(
-      op,
+      "gn2_" + lib_name, "binary_g", in_t, out_t, op, 2, "uint");
-      4);
+  kernel_source += get_template_definition(
+      "gn4large_" + lib_name, "binary_g", in_t, out_t, op, 4);
 }
 
 MTL::ComputePipelineState* get_binary_kernel(
@@ -104,10 +103,11 @@ MTL::ComputePipelineState* get_binary_kernel(
     const std::string op) {
   std::string lib_name = kernel_name.substr(kernel_name.find("_") + 1);
   auto lib = d.get_library(lib_name, [&]() {
-    std::ostringstream kernel_source;
+    std::string kernel_source;
-    kernel_source << metal::utils() << metal::binary_ops() << metal::binary();
+    kernel_source = metal::utils();
-    add_binary_kernels(lib_name, in_type, out_type, op, kernel_source);
+    concatenate(kernel_source, metal::binary_ops(), metal::binary());
-    return kernel_source.str();
+    append_binary_kernels(lib_name, in_type, out_type, op, kernel_source);
+    return kernel_source;
   });
   return d.get_kernel(kernel_name, lib);
 }
@@ -120,11 +120,10 @@ MTL::ComputePipelineState* get_binary_two_kernel(
     const std::string op) {
   std::string lib_name = kernel_name.substr(kernel_name.find("_") + 1);
   auto lib = d.get_library(lib_name, [&]() {
-    std::ostringstream kernel_source;
+    std::string kernel_source = metal::utils();
-    kernel_source << metal::utils() << metal::binary_ops()
+    concatenate(kernel_source, metal::binary_ops(), metal::binary_two());
-                  << metal::binary_two();
+    append_binary_kernels(lib_name, in_type, out_type, op, kernel_source);
-    add_binary_kernels(lib_name, in_type, out_type, op, kernel_source);
+    return kernel_source;
-    return kernel_source.str();
   });
   return d.get_kernel(kernel_name, lib);
 }
@@ -136,24 +135,29 @@ MTL::ComputePipelineState* get_ternary_kernel(
     const std::string op) {
   std::string lib_name = kernel_name.substr(kernel_name.find("_") + 1);
   auto lib = d.get_library(lib_name, [&]() {
-    std::ostringstream kernel_source;
+    auto t_str = get_type_string(type);
+    std::string kernel_source = metal::utils();
+    concatenate(kernel_source, metal::ternary_ops(), metal::ternary());
     const std::array<std::pair<std::string, std::string>, 5> kernel_types = {{
         {"v", "ternary_v"},
         {"v2", "ternary_v2"},
         {"g1", "ternary_g_nd1"},
-        {"g2", "ternary_g_nd2"},
+        {"g2large", "ternary_g_nd2"},
-        {"g3", "ternary_g_nd3"},
+        {"g3large", "ternary_g_nd3"},
     }};
-    kernel_source << metal::utils() << metal::ternary_ops() << metal::ternary();
     for (auto& [name, func] : kernel_types) {
-      std::string template_def;
+      kernel_source +=
-      template_def = get_template_definition(
+          get_template_definition(name + "_" + lib_name, func, t_str, op);
-          name + "_" + lib_name, func, get_type_string(type), op);
-      kernel_source << template_def;
     }
-    kernel_source << get_template_definition(
+    kernel_source += get_template_definition(
-        "gn4_" + lib_name, "ternary_g", get_type_string(type), op, 4);
+        "g2_" + lib_name, "ternary_g_nd2", t_str, op, "uint");
-    return kernel_source.str();
+    kernel_source += get_template_definition(
+        "g3_" + lib_name, "ternary_g_nd3", t_str, op, "uint");
+    kernel_source += get_template_definition(
+        "gn2_" + lib_name, "ternary_g", t_str, op, 2, "uint");
+    kernel_source += get_template_definition(
+        "gn4large_" + lib_name, "ternary_g", t_str, op, 4);
+    return kernel_source;
   });
   return d.get_kernel(kernel_name, lib);
 }
@@ -165,31 +169,43 @@ MTL::ComputePipelineState* get_copy_kernel(
     const array& out) {
   std::string lib_name = kernel_name.substr(kernel_name.find("_") + 1);
   auto lib = d.get_library(lib_name, [&]() {
-    std::ostringstream kernel_source;
+    std::string kernel_source = metal::utils();
+    kernel_source += metal::copy();
     auto in_type = get_type_string(in.dtype());
     auto out_type = get_type_string(out.dtype());
-    kernel_source << metal::utils() << metal::copy()
+    kernel_source +=
-                  << get_template_definition(
+        get_template_definition("s_" + lib_name, "copy_s", in_type, out_type);
-                         "s_" + lib_name, "copy_s", in_type, out_type)
+    kernel_source +=
-                  << get_template_definition(
+        get_template_definition("v_" + lib_name, "copy_v", in_type, out_type);
-                         "v_" + lib_name, "copy_v", in_type, out_type)
+    kernel_source += get_template_definition(
-                  << get_template_definition(
+        "g1_" + lib_name, "copy_g_nd1", in_type, out_type);
-                         "g1_" + lib_name, "copy_g_nd1", in_type, out_type)
+    kernel_source += get_template_definition(
-                  << get_template_definition(
+        "g2_" + lib_name, "copy_g_nd2", in_type, out_type, "int");
-                         "g2_" + lib_name, "copy_g_nd2", in_type, out_type)
+    kernel_source += get_template_definition(
-                  << get_template_definition(
+        "g3_" + lib_name, "copy_g_nd3", in_type, out_type, "int");
-                         "g3_" + lib_name, "copy_g_nd3", in_type, out_type)
+    kernel_source += get_template_definition(
-                  << get_template_definition(
+        "gn2_" + lib_name, "copy_g", in_type, out_type, 2, "int");
-                         "gn4_" + lib_name, "copy_g", in_type, out_type, 4)
+    kernel_source += get_template_definition(
-                  << get_template_definition(
+        "gg1_" + lib_name, "copy_gg_nd1", in_type, out_type);
-                         "gg1_" + lib_name, "copy_gg_nd1", in_type, out_type)
+    kernel_source += get_template_definition(
-                  << get_template_definition(
+        "gg2_" + lib_name, "copy_gg_nd2", in_type, out_type, "int");
-                         "gg2_" + lib_name, "copy_gg_nd2", in_type, out_type)
+    kernel_source += get_template_definition(
-                  << get_template_definition(
+        "gg3_" + lib_name, "copy_gg_nd3", in_type, out_type, "int");
-                         "gg3_" + lib_name, "copy_gg_nd3", in_type, out_type)
+    kernel_source += get_template_definition(
-                  << get_template_definition(
+        "ggn2_" + lib_name, "copy_gg", in_type, out_type, 2, "int");
-                         "ggn4_" + lib_name, "copy_gg", in_type, out_type, 4);
+    kernel_source += get_template_definition(
-    return kernel_source.str();
+        "g2large_" + lib_name, "copy_g_nd2", in_type, out_type);
+    kernel_source += get_template_definition(
+        "g3large_" + lib_name, "copy_g_nd3", in_type, out_type);
+    kernel_source += get_template_definition(
+        "gn4large_" + lib_name, "copy_g", in_type, out_type, 4);
+    kernel_source += get_template_definition(
+        "gg2large_" + lib_name, "copy_gg_nd2", in_type, out_type);
+    kernel_source += get_template_definition(
+        "gg3large_" + lib_name, "copy_gg_nd3", in_type, out_type);
+    kernel_source += get_template_definition(
+        "ggn4large_" + lib_name, "copy_gg", in_type, out_type, 4);
+    return kernel_source;
   });
   return d.get_kernel(kernel_name, lib);
 }
@@ -321,17 +337,17 @@ MTL::ComputePipelineState* get_reduce_init_kernel(
     const std::string& kernel_name,
     const std::string& func_name,
     const std::string& op_name,
-    const array& out) {
+    const Dtype& out_type) {
   auto lib = d.get_library(kernel_name, [&]() {
-    std::ostringstream kernel_source;
     std::string op_type = op_name;
     op_type[0] = std::toupper(op_name[0]);
-    auto out_type = get_type_string(out.dtype());
+    auto out_t = get_type_string(out_type);
-    std::string op = op_type + "<" + out_type + ">";
+    std::string op = op_type + "<" + out_t + ">";
-    kernel_source << metal::utils() << metal::reduce_utils() << metal::reduce();
+    std::string kernel_source = metal::utils();
-    kernel_source << get_template_definition(
+    kernel_source += metal::reduce_utils();
-        kernel_name, func_name, out_type, op);
+    kernel_source += metal::reduce();
-    return kernel_source.str();
+    kernel_source += get_template_definition(kernel_name, func_name, out_t, op);
+    return kernel_source;
   });
   return d.get_kernel(kernel_name, lib);
 }
@@ -341,30 +357,31 @@ MTL::ComputePipelineState* get_reduce_kernel(
     const std::string& kernel_name,
     const std::string& func_name,
     const std::string& op_name,
-    const array& in,
+    const Dtype& in_type,
-    const array& out,
+    const Dtype& out_type,
+    const std::string& idx_t,
     int ndim /* = -1 */,
     int bm /* = -1 */,
     int bn /* = -1 */) {
   auto lib = d.get_library(kernel_name, [&]() {
     std::string op_type = op_name;
     op_type[0] = std::toupper(op_name[0]);
-    std::ostringstream kernel_source;
+    auto in_t = get_type_string(in_type);
-    auto in_type = get_type_string(in.dtype());
+    auto out_t = get_type_string(out_type);
-    auto out_type = get_type_string(out.dtype());
+    std::string op = op_type + "<" + out_t + ">";
-    std::string op = op_type + "<" + out_type + ">";
+    std::string kernel_source = metal::utils();
-    kernel_source << metal::utils() << metal::reduce_utils() << metal::reduce();
+    concatenate(kernel_source, metal::reduce_utils(), metal::reduce());
     if (bm >= 0) {
-      kernel_source << get_template_definition(
+      kernel_source += get_template_definition(
-          kernel_name, func_name, in_type, out_type, op, ndim, bm, bn);
+          kernel_name, func_name, in_t, out_t, op, idx_t, ndim, bm, bn);
     } else if (ndim >= 0) {
-      kernel_source << get_template_definition(
+      kernel_source += get_template_definition(
-          kernel_name, func_name, in_type, out_type, op, ndim);
+          kernel_name, func_name, in_t, out_t, op, idx_t, ndim);
     } else {
-      kernel_source << get_template_definition(
+      kernel_source += get_template_definition(
-          kernel_name, func_name, in_type, out_type, op);
+          kernel_name, func_name, in_t, out_t, op, idx_t);
     }
-    return kernel_source.str();
+    return kernel_source;
   });
   auto st = d.get_kernel(kernel_name, lib);
   return st;

mlx/backend/metal/kernels.h

@@ -81,15 +81,16 @@ MTL::ComputePipelineState* get_reduce_init_kernel(
     const std::string& kernel_name,
     const std::string& func_name,
     const std::string& op_name,
-    const array& out);
+    const Dtype& out_type);
 
 MTL::ComputePipelineState* get_reduce_kernel(
     metal::Device& d,
     const std::string& kernel_name,
     const std::string& func_name,
     const std::string& op_name,
-    const array& in,
+    const Dtype& in_type,
-    const array& out,
+    const Dtype& out_type,
+    const std::string& idx_t,
     int ndim = -1,
     int bm = -1,
     int bn = -1);

mlx/backend/metal/kernels/CMakeLists.txt

@@ -1,13 +1,27 @@
-set(BASE_HEADERS bf16.h bf16_math.h complex.h defines.h expm1f.h utils.h)
+set(BASE_HEADERS
+    metal_3_1/bf16.h
+    metal_3_0/bf16.h
+    bf16_math.h
+    complex.h
+    defines.h
+    expm1f.h
+    utils.h)
 
 function(build_kernel_base TARGET SRCFILE DEPS)
   set(METAL_FLAGS -Wall -Wextra -fno-fast-math)
   if(MLX_METAL_DEBUG)
     set(METAL_FLAGS ${METAL_FLAGS} -gline-tables-only -frecord-sources)
   endif()
+  if(MLX_METAL_VERSION GREATER_EQUAL 310)
+    set(VERSION_INCLUDES
+        ${PROJECT_SOURCE_DIR}/mlx/backend/metal/kernels/metal_3_1)
+  else()
+    set(VERSION_INCLUDES
+        ${PROJECT_SOURCE_DIR}/mlx/backend/metal/kernels/metal_3_0)
+  endif()
   add_custom_command(
     COMMAND xcrun -sdk macosx metal ${METAL_FLAGS} -c ${SRCFILE}
-            -I${PROJECT_SOURCE_DIR} -o ${TARGET}.air
+            -I${PROJECT_SOURCE_DIR} -I${VERSION_INCLUDES} -o ${TARGET}.air
     DEPENDS ${SRCFILE} ${DEPS} ${BASE_HEADERS}
     OUTPUT ${TARGET}.air
     COMMENT "Building ${TARGET}.air"
@@ -30,9 +44,7 @@ build_kernel(layer_norm)
 build_kernel(random)
 build_kernel(rms_norm)
 build_kernel(rope)
-build_kernel(
+build_kernel(scaled_dot_product_attention sdpa_vector.h)
-  scaled_dot_product_attention scaled_dot_product_attention_params.h
-  sdpa_vector.h steel/defines.h steel/gemm/transforms.h steel/utils.h)
 
 set(STEEL_HEADERS
     steel/defines.h
@@ -54,6 +66,24 @@ set(STEEL_HEADERS
     steel/utils/type_traits.h
     steel/utils/integral_constant.h)
 
+set(STEEL_ATTN_HEADERS
+    steel/defines.h
+    steel/utils.h
+    steel/gemm/gemm.h
+    steel/gemm/mma.h
+    steel/gemm/loader.h
+    steel/gemm/transforms.h
+    steel/utils/type_traits.h
+    steel/utils/integral_constant.h
+    steel/attn/attn.h
+    steel/attn/loader.h
+    steel/attn/mma.h
+    steel/attn/params.h
+    steel/attn/transforms.h
+    steel/attn/kernels/steel_attention.h)
+
+build_kernel(steel/attn/kernels/steel_attention ${STEEL_ATTN_HEADERS})
+
 if(NOT MLX_METAL_JIT)
   build_kernel(arange arange.h)
   build_kernel(binary binary.h binary_ops.h)

mlx/backend/metal/kernels/arange.metal

@@ -1,7 +1,7 @@
 // Copyright © 2023-2024 Apple Inc.
 
 // clang-format off
-#include "mlx/backend/metal/kernels/bf16.h"
+#include "mlx/backend/metal/kernels/utils.h"
 #include "mlx/backend/metal/kernels/arange.h"
 
 #define instantiate_arange(tname, type)                                 \

mlx/backend/metal/kernels/arg_reduce.metal

@@ -75,10 +75,10 @@ template <typename T, typename Op, int N_READS = 4>
     const device T* in [[buffer(0)]],
     device uint32_t* out [[buffer(1)]],
     const constant int* shape [[buffer(2)]],
-    const constant size_t* in_strides [[buffer(3)]],
+    const constant int64_t* in_strides [[buffer(3)]],
-    const constant size_t* out_strides [[buffer(4)]],
+    const constant int64_t* out_strides [[buffer(4)]],
     const constant size_t& ndim [[buffer(5)]],
-    const constant size_t& axis_stride [[buffer(6)]],
+    const constant int64_t& axis_stride [[buffer(6)]],
     const constant size_t& axis_size [[buffer(7)]],
     uint gid [[thread_position_in_grid]],
     uint lid [[thread_position_in_threadgroup]],

mlx/backend/metal/kernels/bf16_math.h

@@ -2,8 +2,6 @@
 
 #pragma once
 
-#include "mlx/backend/metal/kernels/bf16.h"
-
 ///////////////////////////////////////////////////////////////////////////////
 // Metal math for bfloat16
 ///////////////////////////////////////////////////////////////////////////////
@@ -369,18 +367,6 @@ instantiate_metal_math_funcs(
     return static_cast<otype>(__metal_simd_xor(static_cast<ctype>(data)));     \
   }
 
-#if (MLX_METAL_VERSION >= 310) || (__METAL_VERSION__ >= 310)
-
-#define bfloat16_to_uint16(x) as_type<uint16_t>(x)
-#define uint16_to_bfloat16(x) as_type<bfloat16_t>(x)
-
-#else
-
-#define bfloat16_to_uint16(x) x.bits_
-#define uint16_to_bfloat16(x) _MLX_BFloat16(x, _MLX_BFloat16::bits_to_bfloat())
-
-#endif
-
 namespace metal {
 
 instantiate_metal_simd_comm_funcs(

mlx/backend/metal/kernels/binary.h

@@ -43,7 +43,7 @@ template <typename T, typename U, typename Op>
     device U* c,
     uint2 index [[thread_position_in_grid]],
     uint2 grid_dim [[threads_per_grid]]) {
-  size_t offset = index.x + grid_dim.x * size_t(index.y);
+  int64_t offset = index.x + grid_dim.x * int64_t(index.y);
   c[offset] = Op()(a[0], b[offset]);
 }
 
@@ -54,7 +54,7 @@ template <typename T, typename U, typename Op>
     device U* c,
     uint2 index [[thread_position_in_grid]],
     uint2 grid_dim [[threads_per_grid]]) {
-  size_t offset = index.x + grid_dim.x * size_t(index.y);
+  int64_t offset = index.x + grid_dim.x * int64_t(index.y);
   c[offset] = Op()(a[offset], b[0]);
 }
 
@@ -65,72 +65,75 @@ template <typename T, typename U, typename Op>
     device U* c,
     uint2 index [[thread_position_in_grid]],
     uint2 grid_dim [[threads_per_grid]]) {
-  size_t offset = index.x + grid_dim.x * size_t(index.y);
+  int64_t offset = index.x + grid_dim.x * int64_t(index.y);
   c[offset] = Op()(a[offset], b[offset]);
 }
 
-template <typename T, typename U, typename Op>
+template <typename T, typename U, typename Op, typename IdxT = int64_t>
 [[kernel]] void binary_g_nd1(
     device const T* a,
     device const T* b,
     device U* c,
-    constant const size_t& a_stride,
+    constant const int64_t& a_stride,
-    constant const size_t& b_stride,
+    constant const int64_t& b_stride,
     uint index [[thread_position_in_grid]]) {
-  auto a_idx = elem_to_loc_1(index, a_stride);
+  auto a_idx = elem_to_loc_1<IdxT>(index, a_stride);
-  auto b_idx = elem_to_loc_1(index, b_stride);
+  auto b_idx = elem_to_loc_1<IdxT>(index, b_stride);
   c[index] = Op()(a[a_idx], b[b_idx]);
 }
 
-template <typename T, typename U, typename Op>
+template <typename T, typename U, typename Op, typename IdxT = int64_t>
 [[kernel]] void binary_g_nd2(
     device const T* a,
     device const T* b,
     device U* c,
-    constant const size_t a_strides[2],
+    constant const int64_t a_strides[2],
-    constant const size_t b_strides[2],
+    constant const int64_t b_strides[2],
     uint2 index [[thread_position_in_grid]],
     uint2 grid_dim [[threads_per_grid]]) {
-  auto a_idx = elem_to_loc_2(index, a_strides);
+  auto a_idx = elem_to_loc_2<IdxT>(index, a_strides);
-  auto b_idx = elem_to_loc_2(index, b_strides);
+  auto b_idx = elem_to_loc_2<IdxT>(index, b_strides);
-  size_t out_idx = index.x + size_t(grid_dim.x) * index.y;
+  IdxT out_idx = index.x + IdxT(grid_dim.x) * index.y;
   c[out_idx] = Op()(a[a_idx], b[b_idx]);
 }
 
-template <typename T, typename U, typename Op>
+template <typename T, typename U, typename Op, typename IdxT = int64_t>
 [[kernel]] void binary_g_nd3(
     device const T* a,
     device const T* b,
     device U* c,
-    constant const size_t a_strides[3],
+    constant const int64_t a_strides[3],
-    constant const size_t b_strides[3],
+    constant const int64_t b_strides[3],
     uint3 index [[thread_position_in_grid]],
     uint3 grid_dim [[threads_per_grid]]) {
-  auto a_idx = elem_to_loc_3(index, a_strides);
+  auto a_idx = elem_to_loc_3<IdxT>(index, a_strides);
-  auto b_idx = elem_to_loc_3(index, b_strides);
+  auto b_idx = elem_to_loc_3<IdxT>(index, b_strides);
-  size_t out_idx =
+  IdxT out_idx = index.x + grid_dim.x * (index.y + IdxT(grid_dim.y) * index.z);
-      index.x + grid_dim.x * (index.y + size_t(grid_dim.y) * index.z);
   c[out_idx] = Op()(a[a_idx], b[b_idx]);
 }
 
-template <typename T, typename U, typename Op, int N = 1>
+template <
+    typename T,
+    typename U,
+    typename Op,
+    int N = 1,
+    typename IdxT = int64_t>
 [[kernel]] void binary_g(
     device const T* a,
     device const T* b,
     device U* c,
     constant const int* shape,
-    constant const size_t* a_strides,
+    constant const int64_t* a_strides,
-    constant const size_t* b_strides,
+    constant const int64_t* b_strides,
     constant const int& ndim,
     uint3 index [[thread_position_in_grid]],
     uint3 grid_dim [[threads_per_grid]]) {
-  auto idx = elem_to_loc_2_nd(
+  auto idx = elem_to_loc_2_nd<IdxT>(
       {N * index.x, index.y, index.z}, shape, a_strides, b_strides, ndim);
   auto xshape = shape[ndim - 1];
-  size_t out_idx =
+  IdxT out_idx = N * index.x + xshape * (index.y + IdxT(grid_dim.y) * index.z);
-      N * index.x + xshape * (index.y + size_t(grid_dim.y) * index.z);
+  IdxT a_xstride = a_strides[ndim - 1];
-  auto a_xstride = a_strides[ndim - 1];
+  IdxT b_xstride = b_strides[ndim - 1];
-  auto b_xstride = b_strides[ndim - 1];
   for (int i = 0; i < N && (int(N * index.x) + i) < xshape; ++i) {
     c[out_idx++] = Op()(a[idx.x], b[idx.y]);
     idx.x += a_xstride;

mlx/backend/metal/kernels/binary.metal

@@ -9,18 +9,22 @@
 #include "mlx/backend/metal/kernels/binary_ops.h"
 #include "mlx/backend/metal/kernels/binary.h"
 
-#define instantiate_binary_all(op, tname, itype, otype)                  \
+#define instantiate_binary_all(op, tname, itype, otype)                        \
-  instantiate_kernel("ss_" #op #tname, binary_ss, itype, otype, op)      \
+  instantiate_kernel("ss_" #op #tname, binary_ss, itype, otype, op)            \
-  instantiate_kernel("sv_" #op #tname, binary_sv, itype, otype, op)      \
+  instantiate_kernel("sv_" #op #tname, binary_sv, itype, otype, op)            \
-  instantiate_kernel("vs_" #op #tname, binary_vs, itype, otype, op)      \
+  instantiate_kernel("vs_" #op #tname, binary_vs, itype, otype, op)            \
-  instantiate_kernel("vv_" #op #tname, binary_vv, itype, otype, op)      \
+  instantiate_kernel("vv_" #op #tname, binary_vv, itype, otype, op)            \
-  instantiate_kernel("sv2_" #op #tname, binary_sv2, itype, otype, op)    \
+  instantiate_kernel("sv2_" #op #tname, binary_sv2, itype, otype, op)          \
-  instantiate_kernel("vs2_" #op #tname, binary_vs2, itype, otype, op)    \
+  instantiate_kernel("vs2_" #op #tname, binary_vs2, itype, otype, op)          \
-  instantiate_kernel("vv2_" #op #tname, binary_vv2, itype, otype, op)    \
+  instantiate_kernel("vv2_" #op #tname, binary_vv2, itype, otype, op)          \
-  instantiate_kernel("gn4_" #op #tname, binary_g, itype, otype, op, 4)   \
+  instantiate_kernel("gn2_" #op #tname, binary_g, itype, otype, op, 2, uint)   \
-  instantiate_kernel("g1_" #op #tname, binary_g_nd1, itype, otype, op)   \
+  instantiate_kernel("gn4large_" #op #tname, binary_g, itype, otype, op, 4)    \
-  instantiate_kernel("g2_" #op #tname, binary_g_nd2, itype, otype, op)   \
+  instantiate_kernel("g1_" #op #tname, binary_g_nd1, itype, otype, op, uint)   \
-  instantiate_kernel("g3_" #op #tname, binary_g_nd3, itype, otype, op)   \
+  instantiate_kernel("g1large_" #op #tname, binary_g_nd1, itype, otype, op)    \
+  instantiate_kernel("g2_" #op #tname, binary_g_nd2, itype, otype, op, uint)   \
+  instantiate_kernel("g2large_" #op #tname, binary_g_nd2, itype, otype, op)    \
+  instantiate_kernel("g3_" #op #tname, binary_g_nd3, itype, otype, op, uint)   \
+  instantiate_kernel("g3large_" #op #tname, binary_g_nd3, itype, otype, op)
 
 #define instantiate_binary_integer(op)                   \
   instantiate_binary_all(op, uint8, uint8_t, uint8_t)    \

mlx/backend/metal/kernels/binary_two.h

@@ -56,7 +56,7 @@ template <typename T, typename U, typename Op>
     device U* d,
     uint2 index [[thread_position_in_grid]],
     uint2 grid_dim [[threads_per_grid]]) {
-  size_t offset = index.x + grid_dim.x * size_t(index.y);
+  auto offset = index.x + grid_dim.x * int64_t(index.y);
   auto out = Op()(a[0], b[offset]);
   c[offset] = out[0];
   d[offset] = out[1];
@@ -70,7 +70,7 @@ template <typename T, typename U, typename Op>
     device U* d,
     uint2 index [[thread_position_in_grid]],
     uint2 grid_dim [[threads_per_grid]]) {
-  size_t offset = index.x + grid_dim.x * size_t(index.y);
+  auto offset = index.x + grid_dim.x * int64_t(index.y);
   auto out = Op()(a[offset], b[0]);
   c[offset] = out[0];
   d[offset] = out[1];
@@ -84,84 +84,87 @@ template <typename T, typename U, typename Op>
     device U* d,
     uint2 index [[thread_position_in_grid]],
     uint2 grid_dim [[threads_per_grid]]) {
-  size_t offset = index.x + grid_dim.x * size_t(index.y);
+  auto offset = index.x + grid_dim.x * int64_t(index.y);
   auto out = Op()(a[offset], b[offset]);
   c[offset] = out[0];
   d[offset] = out[1];
 }
 
-template <typename T, typename U, typename Op>
+template <typename T, typename U, typename Op, typename IdxT = int64_t>
 [[kernel]] void binary_g_nd1(
     device const T* a,
     device const T* b,
     device U* c,
     device U* d,
-    constant const size_t& a_stride,
+    constant const int64_t& a_stride,
-    constant const size_t& b_stride,
+    constant const int64_t& b_stride,
     uint index [[thread_position_in_grid]]) {
-  auto a_idx = elem_to_loc_1(index, a_stride);
+  auto a_idx = elem_to_loc_1<IdxT>(index, a_stride);
-  auto b_idx = elem_to_loc_1(index, b_stride);
+  auto b_idx = elem_to_loc_1<IdxT>(index, b_stride);
   auto out = Op()(a[a_idx], b[b_idx]);
   c[index] = out[0];
   d[index] = out[1];
 }
 
-template <typename T, typename U, typename Op>
+template <typename T, typename U, typename Op, typename IdxT = int64_t>
 [[kernel]] void binary_g_nd2(
     device const T* a,
     device const T* b,
     device U* c,
     device U* d,
-    constant const size_t a_strides[2],
+    constant const int64_t a_strides[2],
-    constant const size_t b_strides[2],
+    constant const int64_t b_strides[2],
     uint2 index [[thread_position_in_grid]],
     uint2 grid_dim [[threads_per_grid]]) {
-  auto a_idx = elem_to_loc_2(index, a_strides);
+  auto a_idx = elem_to_loc_2<IdxT>(index, a_strides);
-  auto b_idx = elem_to_loc_2(index, b_strides);
+  auto b_idx = elem_to_loc_2<IdxT>(index, b_strides);
-  size_t out_idx = index.x + size_t(grid_dim.x) * index.y;
+  IdxT out_idx = index.x + IdxT(grid_dim.x) * index.y;
   auto out = Op()(a[a_idx], b[b_idx]);
   c[out_idx] = out[0];
   d[out_idx] = out[1];
 }
 
-template <typename T, typename U, typename Op>
+template <typename T, typename U, typename Op, typename IdxT = int64_t>
 [[kernel]] void binary_g_nd3(
     device const T* a,
     device const T* b,
     device U* c,
     device U* d,
-    constant const size_t a_strides[3],
+    constant const int64_t a_strides[3],
-    constant const size_t b_strides[3],
+    constant const int64_t b_strides[3],
     uint3 index [[thread_position_in_grid]],
     uint3 grid_dim [[threads_per_grid]]) {
-  auto a_idx = elem_to_loc_3(index, a_strides);
+  auto a_idx = elem_to_loc_3<IdxT>(index, a_strides);
-  auto b_idx = elem_to_loc_3(index, b_strides);
+  auto b_idx = elem_to_loc_3<IdxT>(index, b_strides);
-  size_t out_idx =
+  IdxT out_idx = index.x + grid_dim.x * (index.y + IdxT(grid_dim.y) * index.z);
-      index.x + grid_dim.x * (index.y + size_t(grid_dim.y) * index.z);
   auto out = Op()(a[a_idx], b[b_idx]);
   c[out_idx] = out[0];
   d[out_idx] = out[1];
 }
 
-template <typename T, typename U, typename Op, int N = 1>
+template <
+    typename T,
+    typename U,
+    typename Op,
+    int N = 1,
+    typename IdxT = int64_t>
 [[kernel]] void binary_g(
     device const T* a,
     device const T* b,
     device U* c,
     device U* d,
     constant const int* shape,
-    constant const size_t* a_strides,
+    constant const int64_t* a_strides,
-    constant const size_t* b_strides,
+    constant const int64_t* b_strides,
     constant const int& ndim,
     uint3 index [[thread_position_in_grid]],
     uint3 grid_dim [[threads_per_grid]]) {
-  auto idx = elem_to_loc_2_nd(
+  auto idx = elem_to_loc_2_nd<IdxT>(
       {N * index.x, index.y, index.z}, shape, a_strides, b_strides, ndim);
   auto xshape = shape[ndim - 1];
-  size_t out_idx =
+  IdxT out_idx = N * index.x + xshape * (index.y + IdxT(grid_dim.y) * index.z);
-      N * index.x + xshape * (index.y + size_t(grid_dim.y) * index.z);
+  IdxT a_xstride = a_strides[ndim - 1];
-  auto a_xstride = a_strides[ndim - 1];
+  IdxT b_xstride = b_strides[ndim - 1];
-  auto b_xstride = b_strides[ndim - 1];
   for (int i = 0; i < N && (int(N * index.x) + i) < xshape; ++i) {
     auto out = Op()(a[idx.x], b[idx.y]);
     c[out_idx] = out[0];

mlx/backend/metal/kernels/binary_two.metal

@@ -7,18 +7,22 @@
 #include "mlx/backend/metal/kernels/binary_ops.h"
 #include "mlx/backend/metal/kernels/binary_two.h"
 
-#define instantiate_binary_all(op, tname, itype, otype)                  \
+#define instantiate_binary_all(op, tname, itype, otype)                      \
-  instantiate_kernel("ss_" #op #tname, binary_ss, itype, otype, op)      \
+  instantiate_kernel("ss_" #op #tname, binary_ss, itype, otype, op)          \
-  instantiate_kernel("sv_" #op #tname, binary_sv, itype, otype, op)      \
+  instantiate_kernel("sv_" #op #tname, binary_sv, itype, otype, op)          \
-  instantiate_kernel("vs_" #op #tname, binary_vs, itype, otype, op)      \
+  instantiate_kernel("vs_" #op #tname, binary_vs, itype, otype, op)          \
-  instantiate_kernel("vv_" #op #tname, binary_vv, itype, otype, op)      \
+  instantiate_kernel("vv_" #op #tname, binary_vv, itype, otype, op)          \
-  instantiate_kernel("sv2_" #op #tname, binary_sv2, itype, otype, op)    \
+  instantiate_kernel("sv2_" #op #tname, binary_sv2, itype, otype, op)        \
-  instantiate_kernel("vs2_" #op #tname, binary_vs2, itype, otype, op)    \
+  instantiate_kernel("vs2_" #op #tname, binary_vs2, itype, otype, op)        \
-  instantiate_kernel("vv2_" #op #tname, binary_vv2, itype, otype, op)    \
+  instantiate_kernel("vv2_" #op #tname, binary_vv2, itype, otype, op)        \
-  instantiate_kernel("gn4_" #op #tname, binary_g, itype, otype, op, 4)   \
+  instantiate_kernel("gn2_" #op #tname, binary_g, itype, otype, op, 2, uint) \
-  instantiate_kernel("g1_" #op #tname, binary_g_nd1, itype, otype, op)   \
+  instantiate_kernel("gn4large_" #op #tname, binary_g, itype, otype, op, 4)  \
-  instantiate_kernel("g2_" #op #tname, binary_g_nd2, itype, otype, op)   \
+  instantiate_kernel("g1_" #op #tname, binary_g_nd1, itype, otype, op, uint) \
-  instantiate_kernel("g3_" #op #tname, binary_g_nd3, itype, otype, op)   \
+  instantiate_kernel("g2_" #op #tname, binary_g_nd2, itype, otype, op, uint) \
+  instantiate_kernel("g3_" #op #tname, binary_g_nd3, itype, otype, op, uint) \
+  instantiate_kernel("g1large_" #op #tname, binary_g_nd1, itype, otype, op)  \
+  instantiate_kernel("g2large_" #op #tname, binary_g_nd2, itype, otype, op)  \
+  instantiate_kernel("g3large_" #op #tname, binary_g_nd3, itype, otype, op)
 
 #define instantiate_binary_float(op)                \
   instantiate_binary_all(op, float16, half, half)   \

mlx/backend/metal/kernels/conv.metal

@@ -4,8 +4,8 @@
 #include <metal_simdgroup_matrix>
 #include <metal_stdlib>
 
-#include "mlx/backend/metal/kernels/bf16.h"
 #include "mlx/backend/metal/kernels/steel/conv/params.h"
+#include "mlx/backend/metal/kernels/utils.h"
 
 #define MLX_MTL_CONST static constant constexpr const
 

mlx/backend/metal/kernels/copy.h

@@ -22,7 +22,7 @@ template <typename T, typename U>
     device U* dst [[buffer(1)]],
     uint2 index [[thread_position_in_grid]],
     uint2 grid_dim [[threads_per_grid]]) {
-  size_t offset = index.x + grid_dim.x * size_t(index.y);
+  auto offset = index.x + grid_dim.x * int64_t(index.y);
   dst[offset] = static_cast<U>(src[0]);
 }
 
@@ -32,46 +32,46 @@ template <typename T, typename U>
     device U* dst [[buffer(1)]],
     uint2 index [[thread_position_in_grid]],
     uint2 grid_dim [[threads_per_grid]]) {
-  size_t offset = index.x + grid_dim.x * size_t(index.y);
+  auto offset = index.x + grid_dim.x * int64_t(index.y);
   dst[offset] = static_cast<U>(src[offset]);
 }
 
-template <typename T, typename U>
+template <typename T, typename U, typename IdxT = int64_t>
 [[kernel]] void copy_g_nd1(
     device const T* src [[buffer(0)]],
     device U* dst [[buffer(1)]],
     constant const int64_t& src_stride [[buffer(3)]],
     uint index [[thread_position_in_grid]]) {
-  auto src_idx = elem_to_loc_1(index, src_stride);
+  auto src_idx = elem_to_loc_1<IdxT>(index, src_stride);
   dst[index] = static_cast<U>(src[src_idx]);
 }
 
-template <typename T, typename U>
+template <typename T, typename U, typename IdxT = int64_t>
 [[kernel]] void copy_g_nd2(
     device const T* src [[buffer(0)]],
     device U* dst [[buffer(1)]],
     constant const int64_t* src_strides [[buffer(3)]],
     uint2 index [[thread_position_in_grid]],
     uint2 grid_dim [[threads_per_grid]]) {
-  auto src_idx = elem_to_loc_2(index, src_strides);
+  auto src_idx = elem_to_loc_2<IdxT>(index, src_strides);
-  int64_t dst_idx = index.x + (int64_t)grid_dim.x * index.y;
+  IdxT dst_idx = index.x + IdxT(grid_dim.x) * index.y;
   dst[dst_idx] = static_cast<U>(src[src_idx]);
 }
 
-template <typename T, typename U>
+template <typename T, typename U, typename IdxT = int64_t>
 [[kernel]] void copy_g_nd3(
     device const T* src [[buffer(0)]],
     device U* dst [[buffer(1)]],
     constant const int64_t* src_strides [[buffer(3)]],
     uint3 index [[thread_position_in_grid]],
     uint3 grid_dim [[threads_per_grid]]) {
-  auto src_idx = elem_to_loc_3(index, src_strides);
+  auto src_idx = elem_to_loc_3<IdxT>(index, src_strides);
-  int64_t dst_idx =
+  IdxT dst_idx =
-      index.x + (int64_t)grid_dim.x * (index.y + (int64_t)grid_dim.y * index.z);
+      index.x + IdxT(grid_dim.x) * (index.y + IdxT(grid_dim.y) * index.z);
   dst[dst_idx] = static_cast<U>(src[src_idx]);
 }
 
-template <typename T, typename U, int N = 1>
+template <typename T, typename U, int N = 1, typename IdxT = int64_t>
 [[kernel]] void copy_g(
     device const T* src [[buffer(0)]],
     device U* dst [[buffer(1)]],
@@ -80,17 +80,16 @@ template <typename T, typename U, int N = 1>
     constant const int& ndim [[buffer(5)]],
     uint3 index [[thread_position_in_grid]],
     uint3 grid_dim [[threads_per_grid]]) {
-  auto src_idx = elem_to_loc(
+  auto src_idx = elem_to_loc<IdxT>(
       {N * index.x, index.y, index.z}, src_shape, src_strides, ndim);
   if (N == 1) {
-    int64_t dst_idx =
+    IdxT dst_idx =
-        index.x + grid_dim.x * (index.y + int64_t(grid_dim.y) * index.z);
+        index.x + grid_dim.x * (index.y + IdxT(grid_dim.y) * index.z);
     dst[dst_idx] = static_cast<U>(src[src_idx]);
     return;
   }
   auto xshape = src_shape[ndim - 1];
-  int64_t dst_idx =
+  IdxT dst_idx = N * index.x + xshape * (index.y + IdxT(grid_dim.y) * index.z);
-      N * index.x + xshape * (index.y + int64_t(grid_dim.y) * index.z);
   auto src_xstride = src_strides[ndim - 1];
   for (int i = 0; i < N && (int(N * index.x) + i) < xshape; ++i) {
     dst[dst_idx + i] = static_cast<U>(src[src_idx]);
@@ -98,43 +97,43 @@ template <typename T, typename U, int N = 1>
   }
 }
 
-template <typename T, typename U>
+template <typename T, typename U, typename IdxT = int64_t>
 [[kernel]] void copy_gg_nd1(
     device const T* src [[buffer(0)]],
     device U* dst [[buffer(1)]],
     constant const int64_t& src_stride [[buffer(3)]],
     constant const int64_t& dst_stride [[buffer(4)]],
     uint index [[thread_position_in_grid]]) {
-  auto src_idx = elem_to_loc_1(index, src_stride);
+  auto src_idx = elem_to_loc_1<IdxT>(index, src_stride);
-  auto dst_idx = elem_to_loc_1(index, dst_stride);
+  auto dst_idx = elem_to_loc_1<IdxT>(index, dst_stride);
   dst[dst_idx] = static_cast<U>(src[src_idx]);
 }
 
-template <typename T, typename U>
+template <typename T, typename U, typename IdxT = int64_t>
 [[kernel]] void copy_gg_nd2(
     device const T* src [[buffer(0)]],
     device U* dst [[buffer(1)]],
     constant const int64_t* src_strides [[buffer(3)]],
     constant const int64_t* dst_strides [[buffer(4)]],
     uint2 index [[thread_position_in_grid]]) {
-  auto src_idx = elem_to_loc_2(index, src_strides);
+  auto src_idx = elem_to_loc_2<IdxT>(index, src_strides);
-  auto dst_idx = elem_to_loc_2(index, dst_strides);
+  auto dst_idx = elem_to_loc_2<IdxT>(index, dst_strides);
   dst[dst_idx] = static_cast<U>(src[src_idx]);
 }
 
-template <typename T, typename U>
+template <typename T, typename U, typename IdxT = int64_t>
 [[kernel]] void copy_gg_nd3(
     device const T* src [[buffer(0)]],
     device U* dst [[buffer(1)]],
     constant const int64_t* src_strides [[buffer(3)]],
     constant const int64_t* dst_strides [[buffer(4)]],
     uint3 index [[thread_position_in_grid]]) {
-  auto src_idx = elem_to_loc_3(index, src_strides);
+  auto src_idx = elem_to_loc_3<IdxT>(index, src_strides);
-  auto dst_idx = elem_to_loc_3(index, dst_strides);
+  auto dst_idx = elem_to_loc_3<IdxT>(index, dst_strides);
   dst[dst_idx] = static_cast<U>(src[src_idx]);
 }
 
-template <typename T, typename U, int N = 1>
+template <typename T, typename U, int N = 1, typename IdxT = int64_t>
 [[kernel]] void copy_gg(
     device const T* src [[buffer(0)]],
     device U* dst [[buffer(1)]],
@@ -143,7 +142,7 @@ template <typename T, typename U, int N = 1>
     constant const int64_t* dst_strides [[buffer(4)]],
     constant const int& ndim [[buffer(5)]],
     uint3 index [[thread_position_in_grid]]) {
-  auto idx = elem_to_loc_2_nd(
+  auto idx = elem_to_loc_2_nd<IdxT>(
       {N * index.x, index.y, index.z},
       src_shape,
       src_strides,
@@ -153,8 +152,8 @@ template <typename T, typename U, int N = 1>
     dst[idx.y] = static_cast<U>(src[idx.x]);
     return;
   }
-  auto src_xstride = src_strides[ndim - 1];
+  IdxT src_xstride = src_strides[ndim - 1];
-  auto dst_xstride = dst_strides[ndim - 1];
+  IdxT dst_xstride = dst_strides[ndim - 1];
   auto xshape = src_shape[ndim - 1];
   for (int i = 0; i < N && (int(N * index.x) + i) < xshape; ++i) {
     dst[idx.y] = static_cast<U>(src[idx.x]);

mlx/backend/metal/kernels/copy.metal

@@ -2,22 +2,29 @@
 
 // clang-format off
 #include "mlx/backend/metal/kernels/utils.h"
-#include "mlx/backend/metal/kernels/bf16.h"
 #include "mlx/backend/metal/kernels/copy.h"
 
-#define instantiate_copy_all(tname, itype, otype)    \
+#define instantiate_copy_all(tname, itype, otype)                       \
-  instantiate_kernel("s_copy" #tname, copy_s, itype, otype) \
+  instantiate_kernel("s_copy" #tname, copy_s, itype, otype)             \
-  instantiate_kernel("v_copy" #tname, copy_v, itype, otype) \
+  instantiate_kernel("v_copy" #tname, copy_v, itype, otype)             \
-  instantiate_kernel("s2_copy" #tname, copy_s2, itype, otype) \
+  instantiate_kernel("s2_copy" #tname, copy_s2, itype, otype)           \
-  instantiate_kernel("v2_copy" #tname, copy_v2, itype, otype) \
+  instantiate_kernel("v2_copy" #tname, copy_v2, itype, otype)           \
-  instantiate_kernel("g1_copy" #tname, copy_g_nd1, itype, otype) \
+  instantiate_kernel("g1_copy" #tname, copy_g_nd1, itype, otype, int)   \
-  instantiate_kernel("g2_copy" #tname, copy_g_nd2, itype, otype) \
+  instantiate_kernel("g2_copy" #tname, copy_g_nd2, itype, otype, int)   \
-  instantiate_kernel("g3_copy" #tname, copy_g_nd3, itype, otype) \
+  instantiate_kernel("g3_copy" #tname, copy_g_nd3, itype, otype, int)   \
-  instantiate_kernel("gg1_copy" #tname, copy_gg_nd1, itype, otype) \
+  instantiate_kernel("gg1_copy" #tname, copy_gg_nd1, itype, otype, int) \
-  instantiate_kernel("gg2_copy" #tname, copy_gg_nd2, itype, otype) \
+  instantiate_kernel("gg2_copy" #tname, copy_gg_nd2, itype, otype, int) \
-  instantiate_kernel("gg3_copy" #tname, copy_gg_nd3, itype, otype) \
+  instantiate_kernel("gg3_copy" #tname, copy_gg_nd3, itype, otype, int) \
-  instantiate_kernel("gn4_copy" #tname, copy_g, itype, otype, 4) \
+  instantiate_kernel("gn2_copy" #tname, copy_g, itype, otype, 2, int)   \
-  instantiate_kernel("ggn4_copy" #tname, copy_gg, itype, otype, 4)
+  instantiate_kernel("ggn2_copy" #tname, copy_gg, itype, otype, 2, int) \
+  instantiate_kernel("g1large_copy" #tname, copy_g_nd1, itype, otype)   \
+  instantiate_kernel("g2large_copy" #tname, copy_g_nd2, itype, otype)   \
+  instantiate_kernel("g3large_copy" #tname, copy_g_nd3, itype, otype)   \
+  instantiate_kernel("gg1large_copy" #tname, copy_gg_nd1, itype, otype) \
+  instantiate_kernel("gg2large_copy" #tname, copy_gg_nd2, itype, otype) \
+  instantiate_kernel("gg3large_copy" #tname, copy_gg_nd3, itype, otype) \
+  instantiate_kernel("gn4large_copy" #tname, copy_g, itype, otype, 4)   \
+  instantiate_kernel("ggn4large_copy" #tname, copy_gg, itype, otype, 4)
 
 #define instantiate_copy_itype(itname, itype)                \
   instantiate_copy_all(itname ##bool_, itype, bool)          \

mlx/backend/metal/kernels/gather.h

@@ -4,30 +4,30 @@
 
 #include "mlx/backend/metal/kernels/indexing.h"
 
-template <typename T, typename IdxT, int NIDX, int IDX_NDIM>
+template <typename T, typename IdxT, int NIDX, int IDX_NDIM, typename LocT>
 METAL_FUNC void gather_impl(
     const device T* src [[buffer(0)]],
     device T* out [[buffer(1)]],
     const constant int* src_shape [[buffer(2)]],
-    const constant size_t* src_strides [[buffer(3)]],
+    const constant int64_t* src_strides [[buffer(3)]],
     const constant size_t& src_ndim [[buffer(4)]],
     const constant int* slice_sizes [[buffer(5)]],
     const constant int* axes [[buffer(6)]],
     const thread Indices<IdxT, NIDX>& indices,
     uint3 index [[thread_position_in_grid]],
     uint3 grid_dim [[threads_per_grid]]) {
-  size_t src_idx = 0;
+  LocT src_idx = 0;
   for (int i = 0; i < NIDX; ++i) {
-    size_t idx_loc;
+    LocT idx_loc;
     if (IDX_NDIM == 0) {
       idx_loc = 0;
     } else if (IDX_NDIM == 1) {
-      idx_loc = index.x * indices.strides[indices.ndim * i];
+      idx_loc = index.x * static_cast<LocT>(indices.strides[indices.ndim * i]);
     } else {
-      idx_loc = index.x * indices.strides[indices.ndim * i];
+      idx_loc = index.x * static_cast<LocT>(indices.strides[indices.ndim * i]);
       idx_loc += indices.row_contiguous[i]
           ? index.y
-          : elem_to_loc(
+          : elem_to_loc<LocT>(
                 index.y,
                 &indices.shapes[indices.ndim * i + 1],
                 &indices.strides[indices.ndim * i + 1],
@@ -35,17 +35,17 @@ METAL_FUNC void gather_impl(
     }
     auto ax = axes[i];
     auto idx_val = offset_neg_idx(indices.buffers[i][idx_loc], src_shape[ax]);
-    src_idx += idx_val * src_strides[ax];
+    src_idx += static_cast<LocT>(idx_val) * static_cast<LocT>(src_strides[ax]);
   }
 
-  auto src_offset = elem_to_loc(index.z, slice_sizes, src_strides, src_ndim);
+  auto src_offset =
+      elem_to_loc<LocT>(index.z, slice_sizes, src_strides, src_ndim);
 
-  size_t out_idx = index.z;
+  LocT out_idx = index.z;
   if (IDX_NDIM == 1) {
-    out_idx += static_cast<size_t>(grid_dim.z) * index.x;
+    out_idx += static_cast<LocT>(grid_dim.z) * index.x;
   } else if (IDX_NDIM >= 2) {
-    out_idx +=
+    out_idx += grid_dim.z * (index.x * static_cast<LocT>(grid_dim.y) + index.y);
-        grid_dim.z * (index.x * static_cast<size_t>(grid_dim.y) + index.y);
   }
   out[out_idx] = src[src_offset + src_idx];
 }

mlx/backend/metal/kernels/gemv.metal

@@ -3,8 +3,6 @@
 #include <metal_simdgroup>
 #include <metal_stdlib>
 
-#include "mlx/backend/metal/kernels/bf16.h"
-#include "mlx/backend/metal/kernels/defines.h"
 #include "mlx/backend/metal/kernels/utils.h"
 
 #include "mlx/backend/metal/kernels/steel/utils.h"
@@ -438,9 +436,9 @@ template <
     const constant float& beta [[buffer(8)]],
     const constant int& batch_ndim [[buffer(9)]],
     const constant int* batch_shape [[buffer(10)]],
-    const constant size_t* vector_batch_stride [[buffer(11)]],
+    const constant int64_t* vector_batch_stride [[buffer(11)]],
-    const constant size_t* matrix_batch_stride [[buffer(12)]],
+    const constant int64_t* matrix_batch_stride [[buffer(12)]],
-    const constant size_t* bias_batch_stride [[buffer(13)]],
+    const constant int64_t* bias_batch_stride [[buffer(13)]],
     const constant int& bias_stride [[buffer(14)]],
     uint3 tid [[threadgroup_position_in_grid]],
     uint3 lid [[thread_position_in_threadgroup]],
@@ -488,31 +486,21 @@ template <
       simd_lid);
 }
 
-#define instantiate_gemv_helper(                                             \
+#define instantiate_gemv_helper(                                      \
-    name, itype, bm, bn, sm, sn, tm, tn, nc, axpby)                          \
+    name, itype, bm, bn, sm, sn, tm, tn, nc, axpby)                   \
-  template [[host_name("gemv_" #name "_bm" #bm "_bn" #bn "_sm" #sm "_sn" #sn \
+  instantiate_kernel(                                                 \
-                       "_tm" #tm "_tn" #tn "_nc" #nc                         \
+      "gemv_" #name "_bm" #bm "_bn" #bn "_sm" #sm "_sn" #sn "_tm" #tm \
-                       "_axpby" #axpby)]] [[kernel]] void                    \
+      "_tn" #tn "_nc" #nc "_axpby" #axpby,                            \
-  gemv<itype, bm, bn, sm, sn, tm, tn, nc, axpby>(                            \
+      gemv,                                                           \
-      const device itype* mat [[buffer(0)]],                                 \
+      itype,                                                          \
-      const device itype* in_vec [[buffer(1)]],                              \
+      bm,                                                             \
-      const device itype* bias [[buffer(2)]],                                \
+      bn,                                                             \
-      device itype* out_vec [[buffer(3)]],                                   \
+      sm,                                                             \
-      const constant int& in_vec_size [[buffer(4)]],                         \
+      sn,                                                             \
-      const constant int& out_vec_size [[buffer(5)]],                        \
+      tm,                                                             \
-      const constant int& marix_ld [[buffer(6)]],                            \
+      tn,                                                             \
-      const constant float& alpha [[buffer(7)]],                             \
+      nc,                                                             \
-      const constant float& beta [[buffer(8)]],                              \
+      axpby)
-      const constant int& batch_ndim [[buffer(9)]],                          \
-      const constant int* batch_shape [[buffer(10)]],                        \
-      const constant size_t* vector_batch_stride [[buffer(11)]],             \
-      const constant size_t* matrix_batch_stride [[buffer(12)]],             \
-      const constant size_t* bias_batch_stride [[buffer(13)]],               \
-      const constant int& bias_stride [[buffer(14)]],                        \
-      uint3 tid [[threadgroup_position_in_grid]],                            \
-      uint3 lid [[thread_position_in_threadgroup]],                          \
-      uint simd_gid [[simdgroup_index_in_threadgroup]],                      \
-      uint simd_lid [[thread_index_in_simdgroup]]);
 
 // clang-format off
 #define instantiate_gemv(name, itype, bm, bn, tm, tn)              \
@@ -551,13 +539,13 @@ template <
     const constant float& beta [[buffer(8)]],
     const constant int& batch_ndim [[buffer(9)]],
     const constant int* batch_shape [[buffer(10)]],
-    const constant size_t* index_batch_strides [[buffer(11)]],
+    const constant int64_t* index_batch_strides [[buffer(11)]],
     const constant int& vector_batch_ndim [[buffer(12)]],
     const constant int* vector_batch_shape [[buffer(13)]],
-    const constant size_t* vector_batch_stride [[buffer(14)]],
+    const constant int64_t* vector_batch_stride [[buffer(14)]],
     const constant int& matrix_batch_ndim [[buffer(15)]],
     const constant int* matrix_batch_shape [[buffer(16)]],
-    const constant size_t* matrix_batch_stride [[buffer(17)]],
+    const constant int64_t* matrix_batch_stride [[buffer(17)]],
     const constant uint32_t* vec_indices [[buffer(18)]],
     const constant uint32_t* mat_indices [[buffer(19)]],
     uint3 tid [[threadgroup_position_in_grid]],
@@ -573,8 +561,8 @@ template <
 
   // Update batch offsets
   if (batch_ndim > 1) {
-    const constant size_t* veci_bstrides = index_batch_strides;
+    const constant auto* veci_bstrides = index_batch_strides;
-    const constant size_t* mati_bstrides = index_batch_strides + batch_ndim;
+    const constant auto* mati_bstrides = index_batch_strides + batch_ndim;
 
     ulong2 batch_offsets = elem_to_loc_broadcast(
         tid.z, batch_shape, veci_bstrides, mati_bstrides, batch_ndim);
@@ -621,37 +609,14 @@ template <
       simd_lid);
 }
 
-#define instantiate_gemv_bs_helper(nm, itype, bm, bn, sm, sn, tm, tn)   \
-  template [[host_name("gemv_gather_" #nm "_bm" #bm "_bn" #bn "_sm" #sm \
-                       "_sn" #sn "_tm" #tm "_tn" #tn)]] [[kernel]] void \
-  gemv_gather<itype, bm, bn, sm, sn, tm, tn>(                           \
-      const device itype* mat [[buffer(0)]],                            \
-      const device itype* in_vec [[buffer(1)]],                         \
-      const device itype* bias [[buffer(2)]],                           \
-      device itype* out_vec [[buffer(3)]],                              \
-      const constant int& in_vec_size [[buffer(4)]],                    \
-      const constant int& out_vec_size [[buffer(5)]],                   \
-      const constant int& marix_ld [[buffer(6)]],                       \
-      const constant float& alpha [[buffer(7)]],                        \
-      const constant float& beta [[buffer(8)]],                         \
-      const constant int& batch_ndim [[buffer(9)]],                     \
-      const constant int* batch_shape [[buffer(10)]],                   \
-      const constant size_t* index_batch_strides [[buffer(11)]],        \
-      const constant int& vector_batch_ndim [[buffer(12)]],             \
-      const constant int* vector_batch_shape [[buffer(13)]],            \
-      const constant size_t* vector_batch_stride [[buffer(14)]],        \
-      const constant int& matrix_batch_ndim [[buffer(15)]],             \
-      const constant int* matrix_batch_shape [[buffer(16)]],            \
-      const constant size_t* matrix_batch_stride [[buffer(17)]],        \
-      const constant uint32_t* vec_indices [[buffer(18)]],              \
-      const constant uint32_t* mat_indices [[buffer(19)]],              \
-      uint3 tid [[threadgroup_position_in_grid]],                       \
-      uint3 lid [[thread_position_in_threadgroup]],                     \
-      uint simd_gid [[simdgroup_index_in_threadgroup]],                 \
-      uint simd_lid [[thread_index_in_simdgroup]]);
-
 // clang-format off
-#define instantiate_gemv_bs_blocks(name, itype)        \
+#define instantiate_gemv_bs_helper(nm, itype, bm, bn, sm, sn, tm, tn) \
+  instantiate_kernel(                                                 \
+    "gemv_gather_" #nm "_bm" #bm "_bn" #bn "_sm" #sm                  \
+                       "_sn" #sn "_tm" #tm "_tn" #tn,                 \
+    gemv_gather, itype, bm, bn, sm, sn, tm, tn)
+
+#define instantiate_gemv_bs_blocks(name, itype)              \
   instantiate_gemv_bs_helper(name, itype, 4, 1, 1, 32, 1, 4) \
   instantiate_gemv_bs_helper(name, itype, 4, 1, 1, 32, 4, 4) \
   instantiate_gemv_bs_helper(name, itype, 8, 1, 1, 32, 4, 4) // clang-format on
@@ -686,9 +651,9 @@ template <
     const constant float& beta [[buffer(8)]],
     const constant int& batch_ndim [[buffer(9)]],
     const constant int* batch_shape [[buffer(10)]],
-    const constant size_t* vector_batch_stride [[buffer(11)]],
+    const constant int64_t* vector_batch_stride [[buffer(11)]],
-    const constant size_t* matrix_batch_stride [[buffer(12)]],
+    const constant int64_t* matrix_batch_stride [[buffer(12)]],
-    const constant size_t* bias_batch_stride [[buffer(13)]],
+    const constant int64_t* bias_batch_stride [[buffer(13)]],
     const constant int& bias_stride [[buffer(14)]],
     uint3 tid [[threadgroup_position_in_grid]],
     uint3 lid [[thread_position_in_threadgroup]],
@@ -736,33 +701,14 @@ template <
       simd_lid);
 }
 
-#define instantiate_gemv_t_helper(                                             \
-    name, itype, bm, bn, sm, sn, tm, tn, nc, axpby)                            \
-  template [[host_name("gemv_t_" #name "_bm" #bm "_bn" #bn "_sm" #sm "_sn" #sn \
-                       "_tm" #tm "_tn" #tn "_nc" #nc                           \
-                       "_axpby" #axpby)]] [[kernel]] void                      \
-  gemv_t<itype, bm, bn, sm, sn, tm, tn, nc, axpby>(                            \
-      const device itype* mat [[buffer(0)]],                                   \
-      const device itype* in_vec [[buffer(1)]],                                \
-      const device itype* bias [[buffer(2)]],                                  \
-      device itype* out_vec [[buffer(3)]],                                     \
-      const constant int& in_vec_size [[buffer(4)]],                           \
-      const constant int& out_vec_size [[buffer(5)]],                          \
-      const constant int& marix_ld [[buffer(6)]],                              \
-      const constant float& alpha [[buffer(7)]],                               \
-      const constant float& beta [[buffer(8)]],                                \
-      const constant int& batch_ndim [[buffer(9)]],                            \
-      const constant int* batch_shape [[buffer(10)]],                          \
-      const constant size_t* vector_batch_stride [[buffer(11)]],               \
-      const constant size_t* matrix_batch_stride [[buffer(12)]],               \
-      const constant size_t* bias_batch_stride [[buffer(13)]],                 \
-      const constant int& bias_stride [[buffer(14)]],                          \
-      uint3 tid [[threadgroup_position_in_grid]],                              \
-      uint3 lid [[thread_position_in_threadgroup]],                            \
-      uint simd_gid [[simdgroup_index_in_threadgroup]],                        \
-      uint simd_lid [[thread_index_in_simdgroup]]);
-
 // clang-format off
+#define instantiate_gemv_t_helper(                          \
+    name, itype, bm, bn, sm, sn, tm, tn, nc, axpby)         \
+  instantiate_kernel(                                       \
+    "gemv_t_" #name "_bm" #bm "_bn" #bn "_sm" #sm "_sn" #sn \
+       "_tm" #tm "_tn" #tn "_nc" #nc "_axpby" #axpby,       \
+  gemv_t, itype, bm, bn, sm, sn, tm, tn, nc, axpby)
+
 #define instantiate_gemv_t(name, itype, bm, bn, sm, sn, tm, tn)        \
   instantiate_gemv_t_helper(name, itype, bm, bn, sm, sn, tm, tn, 0, 0) \
   instantiate_gemv_t_helper(name, itype, bm, bn, sm, sn, tm, tn, 0, 1) \
@@ -802,13 +748,13 @@ template <
     const constant float& beta [[buffer(8)]],
     const constant int& batch_ndim [[buffer(9)]],
     const constant int* batch_shape [[buffer(10)]],
-    const constant size_t* index_batch_strides [[buffer(11)]],
+    const constant int64_t* index_batch_strides [[buffer(11)]],
     const constant int& vector_batch_ndim [[buffer(12)]],
     const constant int* vector_batch_shape [[buffer(13)]],
-    const constant size_t* vector_batch_stride [[buffer(14)]],
+    const constant int64_t* vector_batch_stride [[buffer(14)]],
     const constant int& matrix_batch_ndim [[buffer(15)]],
     const constant int* matrix_batch_shape [[buffer(16)]],
-    const constant size_t* matrix_batch_stride [[buffer(17)]],
+    const constant int64_t* matrix_batch_stride [[buffer(17)]],
     const constant uint32_t* vec_indices [[buffer(18)]],
     const constant uint32_t* mat_indices [[buffer(19)]],
     uint3 tid [[threadgroup_position_in_grid]],
@@ -824,8 +770,8 @@ template <
 
   // Update batch offsets
   if (batch_ndim > 1) {
-    const constant size_t* veci_bstrides = index_batch_strides;
+    const constant auto* veci_bstrides = index_batch_strides;
-    const constant size_t* mati_bstrides = index_batch_strides + batch_ndim;
+    const constant auto* mati_bstrides = index_batch_strides + batch_ndim;
 
     ulong2 batch_offsets = elem_to_loc_broadcast(
         tid.z, batch_shape, veci_bstrides, mati_bstrides, batch_ndim);
@@ -872,36 +818,14 @@ template <
       simd_lid);
 }
 
-#define instantiate_gemv_t_bs_helper(nm, itype, bm, bn, sm, sn, tm, tn)   \
-  template [[host_name("gemv_t_gather_" #nm "_bm" #bm "_bn" #bn "_sm" #sm \
-                       "_sn" #sn "_tm" #tm "_tn" #tn)]] [[kernel]] void   \
-  gemv_t_gather<itype, bm, bn, sm, sn, tm, tn>(                           \
-      const device itype* mat [[buffer(0)]],                              \
-      const device itype* in_vec [[buffer(1)]],                           \
-      const device itype* bias [[buffer(2)]],                             \
-      device itype* out_vec [[buffer(3)]],                                \
-      const constant int& in_vec_size [[buffer(4)]],                      \
-      const constant int& out_vec_size [[buffer(5)]],                     \
-      const constant int& marix_ld [[buffer(6)]],                         \
-      const constant float& alpha [[buffer(7)]],                          \
-      const constant float& beta [[buffer(8)]],                           \
-      const constant int& batch_ndim [[buffer(9)]],                       \
-      const constant int* batch_shape [[buffer(10)]],                     \
-      const constant size_t* index_batch_strides [[buffer(11)]],          \
-      const constant int& vector_batch_ndim [[buffer(12)]],               \
-      const constant int* vector_batch_shape [[buffer(13)]],              \
-      const constant size_t* vector_batch_stride [[buffer(14)]],          \
-      const constant int& matrix_batch_ndim [[buffer(15)]],               \
-      const constant int* matrix_batch_shape [[buffer(16)]],              \
-      const constant size_t* matrix_batch_stride [[buffer(17)]],          \
-      const constant uint32_t* vec_indices [[buffer(18)]],                \
-      const constant uint32_t* mat_indices [[buffer(19)]],                \
-      uint3 tid [[threadgroup_position_in_grid]],                         \
-      uint3 lid [[thread_position_in_threadgroup]],                       \
-      uint simd_gid [[simdgroup_index_in_threadgroup]],                   \
-      uint simd_lid [[thread_index_in_simdgroup]]);
-
 // clang-format off
+#define instantiate_gemv_t_bs_helper(                  \
+    nm, itype, bm, bn, sm, sn, tm, tn)                 \
+  instantiate_kernel(                                  \
+    "gemv_t_gather_" #nm "_bm" #bm "_bn" #bn "_sm" #sm \
+       "_sn" #sn "_tm" #tm "_tn" #tn,                  \
+  gemv_t_gather, itype, bm, bn, sm, sn, tm, tn)
+
 #define instantiate_gemv_t_bs_blocks(name, itype)              \
   instantiate_gemv_t_bs_helper(name, itype, 1,  2, 8, 4, 4, 1) \
   instantiate_gemv_t_bs_helper(name, itype, 1,  2, 8, 4, 4, 4) \
@@ -912,4 +836,4 @@ template <
 // clang-format off
 instantiate_gemv_t_bs_blocks(float32, float);
 instantiate_gemv_t_bs_blocks(float16, half);
-instantiate_gemv_t_bs_blocks(bfloat16, bfloat16_t); // clang-format on
+instantiate_gemv_t_bs_blocks(bfloat16, bfloat16_t); // clang-format on

mlx/backend/metal/kernels/gemv_masked.h

@@ -642,13 +642,13 @@ template <
     const constant int& marix_ld [[buffer(6)]],
     const constant int& batch_ndim [[buffer(9)]],
     const constant int* batch_shape [[buffer(10)]],
-    const constant size_t* vector_batch_stride [[buffer(11)]],
+    const constant int64_t* vector_batch_stride [[buffer(11)]],
-    const constant size_t* matrix_batch_stride [[buffer(12)]],
+    const constant int64_t* matrix_batch_stride [[buffer(12)]],
     const device out_mask_t* out_mask [[buffer(20)]],
     const device op_mask_t* mat_mask [[buffer(21)]],
     const device op_mask_t* vec_mask [[buffer(22)]],
     const constant int* mask_strides [[buffer(23)]],
-    const constant size_t* mask_batch_strides [[buffer(24)]],
+    const constant int64_t* mask_batch_strides [[buffer(24)]],
     uint3 tid [[threadgroup_position_in_grid]],
     uint3 lid [[thread_position_in_threadgroup]],
     uint simd_gid [[simdgroup_index_in_threadgroup]],
@@ -673,8 +673,8 @@ template <
     }
 
     if (has_operand_mask) {
-      const constant size_t* mask_strides_mat = mask_batch_strides;
+      const constant auto* mask_strides_mat = mask_batch_strides;
-      const constant size_t* mask_strides_vec = mask_strides_mat + batch_ndim;
+      const constant auto* mask_strides_vec = mask_strides_mat + batch_ndim;
 
       ulong2 batch_offsets = elem_to_loc_broadcast(
           tid.z, batch_shape, mask_strides_mat, mask_strides_vec, batch_ndim);
@@ -742,13 +742,13 @@ template <
     const constant int& marix_ld [[buffer(6)]],
     const constant int& batch_ndim [[buffer(9)]],
     const constant int* batch_shape [[buffer(10)]],
-    const constant size_t* vector_batch_stride [[buffer(11)]],
+    const constant int64_t* vector_batch_stride [[buffer(11)]],
-    const constant size_t* matrix_batch_stride [[buffer(12)]],
+    const constant int64_t* matrix_batch_stride [[buffer(12)]],
     const device out_mask_t* out_mask [[buffer(20)]],
     const device op_mask_t* mat_mask [[buffer(21)]],
     const device op_mask_t* vec_mask [[buffer(22)]],
     const constant int* mask_strides [[buffer(23)]],
-    const constant size_t* mask_batch_strides [[buffer(24)]],
+    const constant int64_t* mask_batch_strides [[buffer(24)]],
     uint3 tid [[threadgroup_position_in_grid]],
     uint3 lid [[thread_position_in_threadgroup]],
     uint simd_gid [[simdgroup_index_in_threadgroup]],
@@ -773,8 +773,8 @@ template <
     }
 
     if (has_operand_mask) {
-      const constant size_t* mask_strides_mat = mask_batch_strides;
+      const constant auto* mask_strides_mat = mask_batch_strides;
-      const constant size_t* mask_strides_vec = mask_strides_mat + batch_ndim;
+      const constant auto* mask_strides_vec = mask_strides_mat + batch_ndim;
 
       ulong2 batch_offsets = elem_to_loc_broadcast(
           tid.z, batch_shape, mask_strides_mat, mask_strides_vec, batch_ndim);

mlx/backend/metal/kernels/gemv_masked.metal

@@ -4,37 +4,17 @@
 #include <metal_simdgroup>
 #include <metal_stdlib>
 
-#include "mlx/backend/metal/kernels/bf16.h"
-#include "mlx/backend/metal/kernels/defines.h"
 #include "mlx/backend/metal/kernels/utils.h"
 
 #include "mlx/backend/metal/kernels/gemv_masked.h"
 
 #define instantiate_gemv_helper(                                           \
     outm_n, outm_t, opm_n, opm_t, name, itype, bm, bn, sm, sn, tm, tn, nc) \
-  template [[host_name("gemv_outmask_" #outm_n "_opmask_" #opm_n "_" #name \
+  instantiate_kernel(                                                      \
-                       "_bm" #bm "_bn" #bn "_sm" #sm "_sn" #sn "_tm" #tm   \
+    "gemv_outmask_" #outm_n "_opmask_" #opm_n "_" #name                    \
-                       "_tn" #tn "_nc" #nc)]] [[kernel]] void              \
+      "_bm" #bm "_bn" #bn "_sm" #sm "_sn" #sn "_tm" #tm                    \
-  gemv_masked<itype, outm_t, opm_t, bm, bn, sm, sn, tm, tn, nc>(           \
+      "_tn" #tn "_nc" #nc,                                                 \
-      const device itype* mat [[buffer(0)]],                               \
+  gemv_masked, itype, outm_t, opm_t, bm, bn, sm, sn, tm, tn, nc)
-      const device itype* in_vec [[buffer(1)]],                            \
-      device itype* out_vec [[buffer(3)]],                                 \
-      const constant int& in_vec_size [[buffer(4)]],                       \
-      const constant int& out_vec_size [[buffer(5)]],                      \
-      const constant int& marix_ld [[buffer(6)]],                          \
-      const constant int& batch_ndim [[buffer(9)]],                        \
-      const constant int* batch_shape [[buffer(10)]],                      \
-      const constant size_t* vector_batch_stride [[buffer(11)]],           \
-      const constant size_t* matrix_batch_stride [[buffer(12)]],           \
-      const device outm_t* out_mask [[buffer(20)]],                        \
-      const device opm_t* mat_mask [[buffer(21)]],                         \
-      const device opm_t* vec_mask [[buffer(22)]],                         \
-      const constant int* mask_strides [[buffer(23)]],                     \
-      const constant size_t* mask_batch_strides [[buffer(24)]],            \
-      uint3 tid [[threadgroup_position_in_grid]],                          \
-      uint3 lid [[thread_position_in_threadgroup]],                        \
-      uint simd_gid [[simdgroup_index_in_threadgroup]],                    \
-      uint simd_lid [[thread_index_in_simdgroup]]);
 
 #define instantiate_gemv_base(name, itype, bm, bn, sm, sn, tm, tn, nc) \
   instantiate_gemv_helper(bool_, bool, bool_, bool, name, itype, bm, bn, sm, sn, tm, tn, nc)      \
@@ -63,29 +43,11 @@ instantiate_gemv_blocks(bfloat16, bfloat16_t);
 
 #define instantiate_gemv_t_helper(                                           \
     outm_n, outm_t, opm_n, opm_t, name, itype, bm, bn, sm, sn, tm, tn, nc)   \
-  template [[host_name("gemv_t_outmask_" #outm_n "_opmask_" #opm_n "_" #name \
+  instantiate_kernel(                                                        \
-                       "_bm" #bm "_bn" #bn "_sm" #sm "_sn" #sn "_tm" #tm     \
+    "gemv_t_outmask_" #outm_n "_opmask_" #opm_n "_" #name                    \
-                       "_tn" #tn "_nc" #nc)]] [[kernel]] void                \
+      "_bm" #bm "_bn" #bn "_sm" #sm "_sn" #sn "_tm" #tm                      \
-  gemv_t_masked<itype, outm_t, opm_t, bm, bn, sm, sn, tm, tn, nc>(           \
+      "_tn" #tn "_nc" #nc,                                                   \
-      const device itype* mat [[buffer(0)]],                                 \
+  gemv_t_masked, itype, outm_t, opm_t, bm, bn, sm, sn, tm, tn, nc)
-      const device itype* in_vec [[buffer(1)]],                              \
-      device itype* out_vec [[buffer(3)]],                                   \
-      const constant int& in_vec_size [[buffer(4)]],                         \
-      const constant int& out_vec_size [[buffer(5)]],                        \
-      const constant int& marix_ld [[buffer(6)]],                            \
-      const constant int& batch_ndim [[buffer(9)]],                          \
-      const constant int* batch_shape [[buffer(10)]],                        \
-      const constant size_t* vector_batch_stride [[buffer(11)]],             \
-      const constant size_t* matrix_batch_stride [[buffer(12)]],             \
-      const device outm_t* out_mask [[buffer(20)]],                          \
-      const device opm_t* mat_mask [[buffer(21)]],                           \
-      const device opm_t* vec_mask [[buffer(22)]],                           \
-      const constant int* mask_strides [[buffer(23)]],                       \
-      const constant size_t* mask_batch_strides [[buffer(24)]],              \
-      uint3 tid [[threadgroup_position_in_grid]],                            \
-      uint3 lid [[thread_position_in_threadgroup]],                          \
-      uint simd_gid [[simdgroup_index_in_threadgroup]],                      \
-      uint simd_lid [[thread_index_in_simdgroup]]);
 
 #define instantiate_gemv_t_base(name, itype, bm, bn, sm, sn, tm, tn, nc) \
   instantiate_gemv_t_helper(bool_, bool, bool_, bool, name, itype, bm, bn, sm, sn, tm, tn, nc)      \

mlx/backend/metal/kernels/indexing.h

@@ -8,13 +8,13 @@ template <typename IdxT, int NIDX>
 struct Indices {
   const array<const device IdxT*, NIDX> buffers;
   const constant int* shapes;
-  const constant size_t* strides;
+  const constant int64_t* strides;
   const constant bool* row_contiguous;
   const int ndim;
 };
 
 template <typename IdxT>
-METAL_FUNC size_t offset_neg_idx(IdxT idx, size_t size) {
+METAL_FUNC size_t offset_neg_idx(IdxT idx, int size) {
   if (is_unsigned_v<IdxT>) {
     return idx;
   } else {

mlx/backend/metal/kernels/jit/bf16.h

@@ -0,0 +1,16 @@
+// Copyright © 2024 Apple Inc.
+
+// clang-format off
+#define jit_if #if
+#define jit_else #else
+#define jit_endif #endif
+
+jit_if (__METAL_VERSION__ >= 310)
+
+#include "mlx/backend/metal/kernels/metal_3_1/bf16.h"
+
+jit_else
+
+#include "mlx/backend/metal/kernels/metal_3_0/bf16.h"
+
+jit_endif // clang-format on

mlx/backend/metal/kernels/layer_norm.metal

@@ -3,8 +3,6 @@
 #include <metal_common>
 #include <metal_simdgroup>
 
-#include "mlx/backend/metal/kernels/bf16.h"
-#include "mlx/backend/metal/kernels/defines.h"
 #include "mlx/backend/metal/kernels/utils.h"
 
 using namespace metal;

mlx/backend/metal/kernels/metal_3_0/bf16.h

@@ -6,12 +6,6 @@
 
 using namespace metal;
 
-#if (MLX_METAL_VERSION >= 310) || (__METAL_VERSION__ >= 310)
-
-typedef bfloat bfloat16_t;
-
-#else
-
 /////////////////////////////////////////////////////////////////////////////
 // Helpers
 /////////////////////////////////////////////////////////////////////////////
@@ -311,7 +305,10 @@ METAL_FUNC bool isnan(_MLX_BFloat16 x) {
 } // namespace metal
 
 #pragma METAL internals : disable
+inline uint16_t bfloat16_to_uint16(const bfloat16_t x) {
+  return x.bits_;
+}
 
-#endif
+inline bfloat16_t uint16_to_bfloat16(const uint16_t x) {
-
+  return _MLX_BFloat16(x, _MLX_BFloat16::bits_to_bfloat());
-#include "mlx/backend/metal/kernels/bf16_math.h"
+}

mlx/backend/metal/kernels/metal_3_1/bf16.h

@@ -0,0 +1,16 @@
+// Copyright © 2023 Apple Inc.
+
+#pragma once
+
+#include <metal_stdlib>
+
+using namespace metal;
+
+typedef bfloat bfloat16_t;
+inline uint16_t bfloat16_to_uint16(const bfloat16_t x) {
+  return as_type<uint16_t>(x);
+}
+
+inline bfloat16_t uint16_to_bfloat16(const uint16_t x) {
+  return as_type<bfloat16_t>(x);
+}

mlx/backend/metal/kernels/quantized.metal

@@ -72,7 +72,6 @@
 
 #define instantiate_quantized_all_single(type, group_size, bits) \
   instantiate_quantized(affine_quantize, type, group_size, bits) \
-  instantiate_quantized(affine_quantize_scales_biases, type, group_size, bits) \
   instantiate_quantized(affine_dequantize, type, group_size, bits)     \
   instantiate_quantized(bs_qmv_fast, type, group_size, bits)     \
   instantiate_quantized(bs_qmv, type, group_size, bits)     \
@@ -116,7 +115,9 @@
 
 #define instantiate_quantized_all() \
   instantiate_quantized_groups(2) \
+  instantiate_quantized_groups(3) \
   instantiate_quantized_groups(4) \
+  instantiate_quantized_groups(6) \
   instantiate_quantized_groups(8)
 
 instantiate_quantized_all() // clang-format on

mlx/backend/metal/kernels/random.metal

@@ -71,7 +71,7 @@ rbits threefry2x32_hash(const thread uint2& key, uint2 count) {
     constant const uint& bytes_per_key,
     constant const int& ndim,
     constant const int* key_shape,
-    constant const size_t* key_strides,
+    constant const int64_t* key_strides,
     uint2 grid_dim [[threads_per_grid]],
     uint2 index [[thread_position_in_grid]]) {
   auto kidx = 2 * index.x;

mlx/backend/metal/kernels/reduce.metal

@@ -10,186 +10,156 @@
 #include "mlx/backend/metal/kernels/reduction/ops.h"
 #include "mlx/backend/metal/kernels/reduce.h"
 
-#define instantiate_reduce_helper_floats(inst_f, name, op) \
+#define instantiate_init_reduce(name, tname, type, op) \
-  inst_f(name, float16, half, op)                          \
+  instantiate_kernel("init_reduce_" #name #tname, init_reduce, type, op<type>)
-  inst_f(name, float32, float, op)                         \
-  inst_f(name, bfloat16, bfloat16_t, op)
 
-#define instantiate_reduce_helper_uints(inst_f, name, op)  \
+instantiate_init_reduce(and, bool_, bool, And)
-  inst_f(name, uint8, uint8_t, op)                         \
+instantiate_init_reduce(or, bool_, bool, Or)
-  inst_f(name, uint16, uint16_t, op)                       \
-  inst_f(name, uint32, uint32_t, op)
 
-#define instantiate_reduce_helper_ints(inst_f, name, op) \
+#define instantiate_init_sum_prod(name, op)                 \
-  inst_f(name, int8, int8_t, op)                         \
+  instantiate_init_reduce(name, int32, int32_t, op)         \
-  inst_f(name, int16, int16_t, op)                       \
+  instantiate_init_reduce(name, int64, int64_t, op)         \
-  inst_f(name, int32, int32_t, op)
+  instantiate_init_reduce(name, float16, float16_t, op)     \
+  instantiate_init_reduce(name, bfloat16, bfloat16_t, op)   \
+  instantiate_init_reduce(name, float32, float, op)         \
+  instantiate_init_reduce(name, complex64, complex64_t, op)
 
-#define instantiate_reduce_helper_64b(inst_f, name, op) \
+instantiate_init_sum_prod(sum, Sum)
-  inst_f(name, int64, int64_t, op)                      \
+instantiate_init_sum_prod(prod, Prod)
-  inst_f(name, uint64, uint64_t, op)                    \
-  inst_f(name, complex64, complex64_t, op)
 
-#define instantiate_reduce_helper_types(inst_f, name, op) \
+#define instantiate_init_min_max(name, op)                   \
-  instantiate_reduce_helper_floats(inst_f, name, op)      \
+  instantiate_init_reduce(name, bool_, bool, op)             \
-  instantiate_reduce_helper_uints(inst_f, name, op)       \
+  instantiate_init_reduce(name, int8, int8_t, op)            \
-  instantiate_reduce_helper_ints(inst_f, name, op)
+  instantiate_init_reduce(name, int16, int16_t, op)          \
+  instantiate_init_reduce(name, int32, int32_t, op)          \
+  instantiate_init_reduce(name, int64, int64_t, op)          \
+  instantiate_init_reduce(name, uint8, uint8_t, op)          \
+  instantiate_init_reduce(name, uint16, uint16_t, op)        \
+  instantiate_init_reduce(name, uint32, uint32_t, op)        \
+  instantiate_init_reduce(name, uint64, uint64_t, op)        \
+  instantiate_init_reduce(name, float16, float16_t, op)      \
+  instantiate_init_reduce(name, bfloat16, bfloat16_t, op)    \
+  instantiate_init_reduce(name, float32, float, op)          \
+  instantiate_init_reduce(name, complex64, complex64_t, op)
 
-#define instantiate_reduce_ops(inst_f, type_f) \
+instantiate_init_min_max(min, Min)
-  type_f(inst_f, sum, Sum)                     \
+instantiate_init_min_max(max, Max)
-  type_f(inst_f, prod, Prod)                   \
-  type_f(inst_f, min, Min)                     \
-  type_f(inst_f, max, Max)
-
-// Special case for bool reductions
-#define instantiate_reduce_from_types_helper( \
-    inst_f, name, tname, itype, otype, op)    \
-    inst_f(name##tname, itype, otype, op)
-
-#define instantiate_reduce_from_types(inst_f, name, otype, op)  \
-  instantiate_reduce_from_types_helper(                         \
-    inst_f, name, bool_, bool, otype, op)                       \
-  instantiate_reduce_from_types_helper(                         \
-    inst_f, name, uint8, uint8_t, otype, op)                    \
-  instantiate_reduce_from_types_helper(                         \
-    inst_f, name, uint16, uint16_t, otype, op)                  \
-  instantiate_reduce_from_types_helper(                         \
-    inst_f, name, uint32, uint32_t, otype, op)                  \
-  instantiate_reduce_from_types_helper(                         \
-    inst_f, name, uint64, uint64_t, otype, op)                  \
-  instantiate_reduce_from_types_helper(                         \
-    inst_f, name, int8, int8_t, otype, op)                      \
-  instantiate_reduce_from_types_helper(                         \
-    inst_f, name, int16, int16_t, otype, op)                    \
-  instantiate_reduce_from_types_helper(                         \
-    inst_f, name, int32, int32_t, otype, op)                    \
-  instantiate_reduce_from_types_helper(                         \
-    inst_f, name, int64, int64_t, otype, op)                    \
-  instantiate_reduce_from_types_helper(                         \
-    inst_f, name, float16, half, otype, op)                     \
-  instantiate_reduce_from_types_helper(                         \
-    inst_f,                                                     \
-    name,                                                       \
-    float32,                                                    \
-    float,                                                      \
-    otype,                                                      \
-    op)                                                         \
-  instantiate_reduce_from_types_helper(                         \
-    inst_f,                                                     \
-    name,                                                       \
-    bfloat16,                                                   \
-    bfloat16_t,                                                 \
-    otype,                                                      \
-    op)
-
-#define instantiate_init_reduce(name, otype, op) \
-  instantiate_kernel("init_reduce_" #name,       \
-                     init_reduce,                \
-                     otype, op)
-
-#define instantiate_init_reduce_helper(name, tname, type, op) \
-  instantiate_init_reduce(name##tname, type, op<type>)
-
-instantiate_reduce_ops(instantiate_init_reduce_helper, instantiate_reduce_helper_types)
-instantiate_reduce_ops(instantiate_init_reduce_helper, instantiate_reduce_helper_64b)
-
-instantiate_init_reduce(andbool_, bool, And<bool>)
-instantiate_init_reduce(orbool_, bool, Or<bool>)
 
 #define instantiate_all_reduce(name, itype, otype, op) \
   instantiate_kernel("all_reduce_" #name,              \
                      all_reduce,                       \
                      itype, otype, op)
 
-#define instantiate_same_all_reduce_helper(name, tname, type, op) \
+#define instantiate_col_reduce_small(name, itype, otype, op, dim)          \
-  instantiate_all_reduce(name##tname, type, type, op<type>)
+  instantiate_kernel("col_reduce_small_" #dim "_reduce_" #name,            \
+                     col_reduce_small,                                     \
+                     itype, otype, op, uint, dim)                          \
+  instantiate_kernel("col_reduce_longcolumn_" #dim "_reduce_" #name,       \
+                     col_reduce_longcolumn,                                \
+                     itype, otype, op, uint, dim)                          \
+  instantiate_kernel("col_reduce_small_large_" #dim "_reduce_" #name,      \
+                     col_reduce_small,                                     \
+                     itype, otype, op, int64_t, dim)                       \
+  instantiate_kernel("col_reduce_longcolumn_large_" #dim "_reduce_" #name, \
+                     col_reduce_longcolumn,                                \
+                     itype, otype, op, int64_t, dim)
 
-instantiate_reduce_ops(instantiate_same_all_reduce_helper, instantiate_reduce_helper_types)
+#define instantiate_col_reduce_looped_tile(name, itype, otype, op, dim, bm, bn)        \
-instantiate_reduce_ops(instantiate_same_all_reduce_helper, instantiate_reduce_helper_64b)
+  instantiate_kernel("col_reduce_looped_" #dim "_" #bm "_" #bn "_reduce_" #name,       \
+                     col_reduce_looped,                                                \
+                     itype, otype, op, uint, dim, bm, bn)                              \
+  instantiate_kernel("col_reduce_looped_large_" #dim "_" #bm "_" #bn "_reduce_" #name, \
+                     col_reduce_looped,                                                \
+                     itype, otype, op, int64_t, dim, bm, bn)
 
-instantiate_reduce_from_types(instantiate_all_reduce, and, bool, And<bool>)
+#define instantiate_col_reduce_2pass_tile(name, itype, otype, op, dim, bm, bn)        \
-instantiate_reduce_from_types(instantiate_all_reduce, or, bool, Or<bool>)
+  instantiate_kernel("col_reduce_2pass_" #dim "_" #bm "_" #bn "_reduce_" #name,       \
-
+                     col_reduce_2pass,                                                \
-// special case bool with larger output type
+                     itype, otype, op, uint, dim, bm, bn)                             \
-instantiate_all_reduce(sumbool_, bool, uint32_t, Sum<uint32_t>)
+  instantiate_kernel("col_reduce_2pass_large_" #dim "_" #bm "_" #bn "_reduce_" #name, \
-
+                     col_reduce_2pass,                                                \
-#define instantiate_col_reduce_small(name, itype, otype, op, dim)      \
+                     itype, otype, op, int64_t, dim, bm, bn)
-  instantiate_kernel("col_reduce_small_" #dim "_reduce_" #name,        \
-                     col_reduce_small,                                 \
-                     itype, otype, op, dim)                            \
-  instantiate_kernel("col_reduce_longcolumn_" #dim "_reduce_" #name,   \
-                     col_reduce_longcolumn,                            \
-                     itype, otype, op, dim)
-
-#define instantiate_col_reduce_looped_tile(name, itype, otype, op, dim, bm, bn)  \
-  instantiate_kernel("col_reduce_looped_" #dim "_" #bm "_" #bn "_reduce_" #name, \
-                     col_reduce_looped,                                          \
-                     itype, otype, op, dim, bm, bn)
-
-#define instantiate_col_reduce_2pass_tile(name, itype, otype, op, dim, bm, bn)  \
-  instantiate_kernel("col_reduce_2pass_" #dim "_" #bm "_" #bn "_reduce_" #name, \
-                     col_reduce_2pass,                                          \
-                     itype, otype, op, dim, bm, bn)
 
 #define instantiate_col_reduce_looped(name, itype, otype, op, dim)        \
   instantiate_col_reduce_looped_tile(name, itype, otype, op, dim, 32, 32) \
   instantiate_col_reduce_2pass_tile(name, itype, otype, op, dim, 32, 32)
 
 #define instantiate_col_reduce_general(name, itype, otype, op) \
-  instantiate_col_reduce_small(name, itype, otype, op, 0)      \
   instantiate_col_reduce_small(name, itype, otype, op, 1)      \
   instantiate_col_reduce_small(name, itype, otype, op, 2)      \
-  instantiate_col_reduce_small(name, itype, otype, op, 3)      \
+  instantiate_col_reduce_small(name, itype, otype, op, 5)      \
-  instantiate_col_reduce_small(name, itype, otype, op, 4)      \
-  instantiate_col_reduce_looped(name, itype, otype, op, 0)     \
   instantiate_col_reduce_looped(name, itype, otype, op, 1)     \
   instantiate_col_reduce_looped(name, itype, otype, op, 2)     \
-  instantiate_col_reduce_looped(name, itype, otype, op, 3)     \
+  instantiate_col_reduce_looped(name, itype, otype, op, 5)
-  instantiate_col_reduce_looped(name, itype, otype, op, 4)
 
-#define instantiate_same_col_reduce_helper(name, tname, type, op)  \
+#define instantiate_row_reduce_small(name, itype, otype, op, dim)     \
-  instantiate_col_reduce_general(name##tname, type, type, op<type>)
+  instantiate_kernel("row_reduce_small_" #dim "_reduce_" #name,       \
+                     row_reduce_small,                                \
+                     itype, otype, op, uint, dim)                     \
+  instantiate_kernel("row_reduce_small_large_" #dim "_reduce_" #name, \
+                     row_reduce_small,                                \
+                     itype, otype, op, int64_t, dim)
 
-instantiate_reduce_ops(instantiate_same_col_reduce_helper, instantiate_reduce_helper_types)
+#define instantiate_row_reduce_looped(name, itype, otype, op, dim)       \
-instantiate_reduce_ops(instantiate_same_col_reduce_helper, instantiate_reduce_helper_64b)
+  instantiate_kernel("row_reduce_looped_" #dim "_reduce_" #name,         \
-
+                     row_reduce_looped,                                  \
-instantiate_col_reduce_general(sumbool_, bool, uint32_t, Sum<uint32_t>)
+                     itype, otype, op, uint, dim)                        \
-instantiate_reduce_from_types(instantiate_col_reduce_general, and, bool, And<bool>)
+  instantiate_kernel("row_reduce_looped_large_" #dim "_reduce_" #name,   \
-instantiate_reduce_from_types(instantiate_col_reduce_general, or, bool, Or<bool>)
+                     row_reduce_looped,                                  \
-
+                     itype, otype, op, int64_t, dim)
-#define instantiate_row_reduce_small(name, itype, otype, op, dim) \
-  instantiate_kernel("row_reduce_small_" #dim "_reduce_" #name,   \
-                     row_reduce_small,                            \
-                     itype, otype, op, dim)
-
-#define instantiate_row_reduce_looped(name, itype, otype, op, dim) \
-  instantiate_kernel("row_reduce_looped_" #dim "_reduce_" #name,   \
-                     row_reduce_looped, \
-                     itype, otype, op, dim)
 
 #define instantiate_row_reduce_general(name, itype, otype, op) \
-  instantiate_row_reduce_small(name, itype, otype, op, 0)      \
   instantiate_row_reduce_small(name, itype, otype, op, 1)      \
   instantiate_row_reduce_small(name, itype, otype, op, 2)      \
-  instantiate_row_reduce_small(name, itype, otype, op, 3)      \
+  instantiate_row_reduce_small(name, itype, otype, op, 5)      \
-  instantiate_row_reduce_small(name, itype, otype, op, 4)      \
-  instantiate_row_reduce_looped(name, itype, otype, op, 0)     \
   instantiate_row_reduce_looped(name, itype, otype, op, 1)     \
   instantiate_row_reduce_looped(name, itype, otype, op, 2)     \
-  instantiate_row_reduce_looped(name, itype, otype, op, 3)     \
+  instantiate_row_reduce_looped(name, itype, otype, op, 5)     \
-  instantiate_row_reduce_looped(name, itype, otype, op, 4)     \
   instantiate_kernel("row_reduce_simple_" #name,               \
                      row_reduce_simple,                        \
                      itype, otype, op)
 
-#define instantiate_same_row_reduce_helper(name, tname, type, op)  \
+#define instantiate_reduce_functions(name, tname, itype, otype, op)    \
-  instantiate_row_reduce_general(name##tname, type, type, op<type>)
+  instantiate_all_reduce(name##tname, itype, otype, op<otype>)         \
+  instantiate_row_reduce_general(name##tname, itype, otype, op<otype>) \
+  instantiate_col_reduce_general(name##tname, itype, otype, op<otype>)
 
-instantiate_reduce_ops(instantiate_same_row_reduce_helper, instantiate_reduce_helper_types)
+#define instantiate_and_or(name, op)                           \
-instantiate_reduce_ops(instantiate_same_row_reduce_helper, instantiate_reduce_helper_64b)
+  instantiate_reduce_functions(name, bool_, bool, bool, op)    \
+  instantiate_reduce_functions(name, int16, int16_t, bool, op) \
+  instantiate_reduce_functions(name, int32, int32_t, bool, op) \
+  instantiate_reduce_functions(name, int64, int64_t, bool, op)
 
-instantiate_reduce_from_types(instantiate_row_reduce_general, and, bool, And<bool>)
+instantiate_and_or(and, And)
-instantiate_reduce_from_types(instantiate_row_reduce_general, or, bool, Or<bool>)
+instantiate_and_or(or, Or)
 
-instantiate_row_reduce_general(sumbool_, bool, uint32_t, Sum<uint32_t>)
+#define instantiate_sum_prod(name, op)                                       \
+  instantiate_reduce_functions(name, int8, int8_t, int32_t, op)              \
+  instantiate_reduce_functions(name, int16, int16_t, int32_t, op)            \
+  instantiate_reduce_functions(name, int32, int32_t, int32_t, op)            \
+  instantiate_reduce_functions(name, int64, int64_t, int64_t, op)            \
+  instantiate_reduce_functions(name, float16, float16_t, float16_t, op)      \
+  instantiate_reduce_functions(name, bfloat16, bfloat16_t, bfloat16_t, op)   \
+  instantiate_reduce_functions(name, float32, float, float, op)              \
+  instantiate_reduce_functions(name, complex64, complex64_t, complex64_t, op)
+
+instantiate_sum_prod(sum, Sum)
+instantiate_sum_prod(prod, Prod)
+
+#define instantiate_min_max(name, op)                                        \
+  instantiate_reduce_functions(name, int8, int8_t, int8_t, op)               \
+  instantiate_reduce_functions(name, int16, int16_t, int16_t, op)            \
+  instantiate_reduce_functions(name, int32, int32_t, int32_t, op)            \
+  instantiate_reduce_functions(name, int64, int64_t, int64_t, op)            \
+  instantiate_reduce_functions(name, uint8, uint8_t, uint8_t, op)            \
+  instantiate_reduce_functions(name, uint16, uint16_t, uint16_t, op)         \
+  instantiate_reduce_functions(name, uint32, uint32_t, uint32_t, op)         \
+  instantiate_reduce_functions(name, uint64, uint64_t, uint64_t, op)         \
+  instantiate_reduce_functions(name, float16, float16_t, float16_t, op)      \
+  instantiate_reduce_functions(name, bfloat16, bfloat16_t, bfloat16_t, op)   \
+  instantiate_reduce_functions(name, float32, float, float, op)              \
+  instantiate_reduce_functions(name, complex64, complex64_t, complex64_t, op)
+
+instantiate_min_max(min, Min)
+instantiate_min_max(max, Max)
     // clang-format on

mlx/backend/metal/kernels/reduction/reduce_all.h

@@ -1,6 +1,11 @@
 // Copyright © 2023-2024 Apple Inc.
 
-template <typename T, typename U, typename Op, int N_READS = REDUCE_N_READS>
+template <
+    typename T,
+    typename U,
+    typename Op,
+    typename IdxT = int64_t,
+    int N_READS = REDUCE_N_READS>
 [[kernel]] void all_reduce(
     const device T* in [[buffer(0)]],
     device U* out [[buffer(1)]],
@@ -16,10 +21,10 @@ template <typename T, typename U, typename Op, int N_READS = REDUCE_N_READS>
   threadgroup U shared_vals[simd_size];
 
   U total = Op::init;
-  int64_t start_idx = gid.y * row_size;
+  IdxT start_idx = gid.y * IdxT(row_size);
-  int64_t actual_row =
+  IdxT actual_row =
       (start_idx + row_size <= in_size) ? row_size : in_size - start_idx;
-  int64_t blocks = actual_row / (lsize.x * N_READS);
+  IdxT blocks = actual_row / (lsize.x * N_READS);
   int extra = actual_row - blocks * (lsize.x * N_READS);
   extra -= lid.x * N_READS;
   start_idx += lid.x * N_READS;
@@ -30,7 +35,7 @@ template <typename T, typename U, typename Op, int N_READS = REDUCE_N_READS>
     extra = 0;
   }
 
-  for (int64_t b = 0; b < blocks; b++) {
+  for (IdxT b = 0; b < blocks; b++) {
     for (int i = 0; i < N_READS; i++) {
       total = op(static_cast<U>(in[i]), total);
     }

mlx/backend/metal/kernels/reduction/reduce_col.h

@@ -1,16 +1,16 @@
 // Copyright © 2023-2024 Apple Inc.
 
-template <typename T, typename U, typename Op, int NDIMS>
+template <typename T, typename U, typename Op, typename IdxT, int NDIMS>
 [[kernel]] void col_reduce_small(
     const device T* in [[buffer(0)]],
     device U* out [[buffer(1)]],
     const constant size_t& reduction_size [[buffer(2)]],
-    const constant size_t& reduction_stride [[buffer(3)]],
+    const constant int64_t& reduction_stride [[buffer(3)]],
     const constant int* shape [[buffer(4)]],
-    const constant size_t* strides [[buffer(5)]],
+    const constant int64_t* strides [[buffer(5)]],
     const constant int& ndim [[buffer(6)]],
     const constant int* reduce_shape [[buffer(7)]],
-    const constant size_t* reduce_strides [[buffer(8)]],
+    const constant int64_t* reduce_strides [[buffer(8)]],
     const constant int& reduce_ndim [[buffer(9)]],
     const constant size_t& non_col_reductions [[buffer(10)]],
     uint3 gid [[threadgroup_position_in_grid]],
@@ -19,7 +19,7 @@ template <typename T, typename U, typename Op, int NDIMS>
     uint3 lsize [[threads_per_threadgroup]]) {
   constexpr int n_reads = 4;
   Op op;
-  looped_elem_to_loc<NDIMS> loop;
+  LoopedElemToLoc<NDIMS, IdxT, (NDIMS > 2)> loop(reduce_ndim);
   const device T* row;
 
   U totals[n_reads];
@@ -27,20 +27,20 @@ template <typename T, typename U, typename Op, int NDIMS>
     totals[i] = Op::init;
   }
 
-  size_t column = size_t(gid.x) * lsize.x * n_reads + lid.x * n_reads;
+  IdxT column = IdxT(gid.x) * lsize.x * n_reads + lid.x * n_reads;
   if (column >= reduction_stride) {
     return;
   }
   bool safe = column + n_reads <= reduction_stride;
 
-  size_t out_idx = gid.y + gsize.y * size_t(gid.z);
+  IdxT out_idx = gid.y + gsize.y * IdxT(gid.z);
-  size_t in_idx = elem_to_loc(out_idx, shape, strides, ndim);
+  IdxT in_idx = elem_to_loc<IdxT>(out_idx, shape, strides, ndim);
   in += in_idx + column;
 
-  size_t total_rows = non_col_reductions * reduction_size;
+  IdxT total_rows = IdxT(non_col_reductions) * IdxT(reduction_size);
   loop.next(lid.y, reduce_shape, reduce_strides);
-  for (size_t r = lid.y; r < total_rows; r += lsize.y) {
+  for (IdxT r = lid.y; r < total_rows; r += lsize.y) {
-    row = in + loop.location(r, reduce_shape, reduce_strides, reduce_ndim);
+    row = in + loop.location();
     if (safe) {
       for (int i = 0; i < n_reads; i++) {
         totals[i] = op(static_cast<U>(row[i]), totals[i]);
@@ -80,7 +80,7 @@ template <typename T, typename U, typename Op, int NDIMS>
   }
 
   if (lid.y == 0) {
-    out += out_idx * reduction_stride + column;
+    out += out_idx * IdxT(reduction_stride) + column;
     if (safe) {
       for (int i = 0; i < n_reads; i++) {
         out[i] = totals[i];
@@ -93,17 +93,17 @@ template <typename T, typename U, typename Op, int NDIMS>
   }
 }
 
-template <typename T, typename U, typename Op, int NDIMS>
+template <typename T, typename U, typename Op, typename IdxT, int NDIMS>
 [[kernel]] void col_reduce_longcolumn(
     const device T* in [[buffer(0)]],
     device U* out [[buffer(1)]],
     const constant size_t& reduction_size [[buffer(2)]],
     const constant size_t& reduction_stride [[buffer(3)]],
     const constant int* shape [[buffer(4)]],
-    const constant size_t* strides [[buffer(5)]],
+    const constant int64_t* strides [[buffer(5)]],
     const constant int& ndim [[buffer(6)]],
     const constant int* reduce_shape [[buffer(7)]],
-    const constant size_t* reduce_strides [[buffer(8)]],
+    const constant int64_t* reduce_strides [[buffer(8)]],
     const constant int& reduce_ndim [[buffer(9)]],
     const constant size_t& non_col_reductions [[buffer(10)]],
     const constant size_t& out_size [[buffer(11)]],
@@ -112,19 +112,19 @@ template <typename T, typename U, typename Op, int NDIMS>
     uint3 lid [[thread_position_in_threadgroup]],
     uint3 lsize [[threads_per_threadgroup]]) {
   Op op;
-  looped_elem_to_loc<NDIMS> loop;
+  LoopedElemToLoc<NDIMS, IdxT, (NDIMS > 2)> loop(reduce_ndim);
   const device T* row;
 
-  size_t out_idx = gid.x + gsize.x * size_t(gid.y);
+  IdxT out_idx = gid.x + gsize.x * IdxT(gid.y);
-  size_t in_idx = elem_to_loc(out_idx, shape, strides, ndim);
+  IdxT in_idx = elem_to_loc<IdxT>(out_idx, shape, strides, ndim);
   in += in_idx + lid.x;
 
   U total = Op::init;
-  size_t total_rows = non_col_reductions * reduction_size;
+  IdxT total_rows = IdxT(non_col_reductions) * IdxT(reduction_size);
   loop.next(gid.z * lsize.y + lid.y, reduce_shape, reduce_strides);
-  for (size_t r = gid.z * lsize.y + lid.y; r < total_rows;
+  for (IdxT r = gid.z * lsize.y + lid.y; r < total_rows;
        r += lsize.y * gsize.z) {
-    row = in + loop.location(r, reduce_shape, reduce_strides, reduce_ndim);
+    row = in + loop.location();
     total = op(static_cast<U>(*row), total);
     loop.next(lsize.y * gsize.z, reduce_shape, reduce_strides);
   }
@@ -136,7 +136,8 @@ template <typename T, typename U, typename Op, int NDIMS>
     for (uint i = 1; i < lsize.y; i++) {
       total = op(total, shared_vals[i * lsize.x + lid.x]);
     }
-    out[gid.z * out_size + out_idx * reduction_stride + lid.x] = total;
+    out[gid.z * IdxT(out_size) + out_idx * IdxT(reduction_stride) + lid.x] =
+        total;
   }
 }
 
@@ -151,17 +152,24 @@ template <typename T, typename U, typename Op, int NDIMS>
  *     totals with a loop.
  *  7. Write them to the output
  */
-template <typename T, typename U, typename Op, int NDIMS, int BM, int BN>
+template <
+    typename T,
+    typename U,
+    typename Op,
+    typename IdxT,
+    int NDIMS,
+    int BM,
+    int BN>
 [[kernel]] void col_reduce_looped(
     const device T* in [[buffer(0)]],
     device U* out [[buffer(1)]],
     const constant size_t& reduction_size [[buffer(2)]],
-    const constant size_t& reduction_stride [[buffer(3)]],
+    const constant int64_t& reduction_stride [[buffer(3)]],
     const constant int* shape [[buffer(4)]],
-    const constant size_t* strides [[buffer(5)]],
+    const constant int64_t* strides [[buffer(5)]],
     const constant int& ndim [[buffer(6)]],
     const constant int* reduce_shape [[buffer(7)]],
-    const constant size_t* reduce_strides [[buffer(8)]],
+    const constant int64_t* reduce_strides [[buffer(8)]],
     const constant int& reduce_ndim [[buffer(9)]],
     const constant size_t& non_col_reductions [[buffer(10)]],
     uint3 gid [[threadgroup_position_in_grid]],
@@ -176,7 +184,7 @@ template <typename T, typename U, typename Op, int NDIMS, int BM, int BN>
 
   threadgroup U shared_vals[BN * BM];
   U totals[n_reads];
-  looped_elem_to_loc<NDIMS> loop;
+  LoopedElemToLoc<NDIMS, IdxT, (NDIMS > 2)> loop(reduce_ndim);
   const device T* row;
 
   for (int i = 0; i < n_reads; i++) {
@@ -185,17 +193,17 @@ template <typename T, typename U, typename Op, int NDIMS, int BM, int BN>
 
   short lid = simd_group_id * simd_size + simd_lane_id;
   short2 offset((lid % n_read_blocks) * n_reads, lid / n_read_blocks);
-  size_t column = BN * gid.x + offset.x;
+  IdxT column = BN * gid.x + offset.x;
   bool safe = column + n_reads <= reduction_stride;
 
-  size_t out_idx = gid.y + gsize.y * size_t(gid.z);
+  IdxT out_idx = gid.y + gsize.y * IdxT(gid.z);
-  size_t in_idx = elem_to_loc(out_idx, shape, strides, ndim);
+  IdxT in_idx = elem_to_loc<IdxT>(out_idx, shape, strides, ndim);
   in += in_idx + column;
 
-  size_t total = non_col_reductions * reduction_size;
+  IdxT total = IdxT(non_col_reductions) * IdxT(reduction_size);
   loop.next(offset.y, reduce_shape, reduce_strides);
-  for (size_t r = offset.y; r < total; r += BM) {
+  for (IdxT r = offset.y; r < total; r += BM) {
-    row = in + loop.location(r, reduce_shape, reduce_strides, reduce_ndim);
+    row = in + loop.location();
 
     if (safe) {
       for (int i = 0; i < n_reads; i++) {
@@ -235,8 +243,8 @@ template <typename T, typename U, typename Op, int NDIMS, int BM, int BN>
 
     // Write the output.
     if (simd_lane_id == 0) {
-      size_t out_column = BN * gid.x + out_offset.x;
+      IdxT out_column = BN * gid.x + out_offset.x;
-      out += out_idx * reduction_stride + out_column;
+      out += out_idx * IdxT(reduction_stride) + out_column;
       if (out_column + n_outputs <= reduction_stride) {
         for (int i = 0; i < n_outputs; i++) {
           out[i] = totals[i];
@@ -269,7 +277,7 @@ template <typename T, typename U, typename Op, int NDIMS, int BM, int BN>
 
     // Write the output.
     if (offset.y == 0) {
-      out += out_idx * reduction_stride + column;
+      out += out_idx * IdxT(reduction_stride) + column;
       if (safe) {
         for (int i = 0; i < n_reads; i++) {
           out[i] = totals[i];
@@ -283,17 +291,24 @@ template <typename T, typename U, typename Op, int NDIMS, int BM, int BN>
   }
 }
 
-template <typename T, typename U, typename Op, int NDIMS, int BM, int BN>
+template <
+    typename T,
+    typename U,
+    typename Op,
+    typename IdxT,
+    int NDIMS,
+    int BM,
+    int BN>
 [[kernel]] void col_reduce_2pass(
     const device T* in [[buffer(0)]],
     device U* out [[buffer(1)]],
     const constant size_t& reduction_size [[buffer(2)]],
-    const constant size_t& reduction_stride [[buffer(3)]],
+    const constant int64_t& reduction_stride [[buffer(3)]],
     const constant int* shape [[buffer(4)]],
-    const constant size_t* strides [[buffer(5)]],
+    const constant int64_t* strides [[buffer(5)]],
     const constant int& ndim [[buffer(6)]],
     const constant int* reduce_shape [[buffer(7)]],
-    const constant size_t* reduce_strides [[buffer(8)]],
+    const constant int64_t* reduce_strides [[buffer(8)]],
     const constant int& reduce_ndim [[buffer(9)]],
     const constant size_t& non_col_reductions [[buffer(10)]],
     const constant size_t& out_size [[buffer(11)]],
@@ -312,7 +327,7 @@ template <typename T, typename U, typename Op, int NDIMS, int BM, int BN>
 
   threadgroup U shared_vals[BN * BM];
   U totals[n_reads];
-  looped_elem_to_loc<NDIMS> loop;
+  LoopedElemToLoc<NDIMS, IdxT, (NDIMS > 2)> loop(reduce_ndim);
   const device T* row;
 
   for (int i = 0; i < n_reads; i++) {
@@ -321,20 +336,19 @@ template <typename T, typename U, typename Op, int NDIMS, int BM, int BN>
 
   short lid = simd_group_id * simd_size + simd_lane_id;
   short2 offset((lid % n_read_blocks) * n_reads, lid / n_read_blocks);
-  size_t column = BN * gid.x + offset.x;
+  IdxT column = BN * gid.x + offset.x;
   bool safe = column + n_reads <= reduction_stride;
 
-  size_t full_idx = gid.y + gsize.y * size_t(gid.z);
+  IdxT full_idx = gid.y + gsize.y * IdxT(gid.z);
-  size_t block_idx = full_idx / out_size;
+  IdxT block_idx = full_idx / IdxT(out_size);
-  size_t out_idx = full_idx % out_size;
+  IdxT out_idx = full_idx % IdxT(out_size);
-  size_t in_idx = elem_to_loc(out_idx, shape, strides, ndim);
+  IdxT in_idx = elem_to_loc<IdxT>(out_idx, shape, strides, ndim);
   in += in_idx + column;
 
-  size_t total = non_col_reductions * reduction_size;
+  IdxT total = IdxT(non_col_reductions) * IdxT(reduction_size);
   loop.next(offset.y + block_idx * BM, reduce_shape, reduce_strides);
-  for (size_t r = offset.y + block_idx * BM; r < total;
+  for (IdxT r = offset.y + block_idx * BM; r < total; r += outer_blocks * BM) {
-       r += outer_blocks * BM) {
+    row = in + loop.location();
-    row = in + loop.location(r, reduce_shape, reduce_strides, reduce_ndim);
 
     if (safe) {
       for (int i = 0; i < n_reads; i++) {
@@ -369,8 +383,8 @@ template <typename T, typename U, typename Op, int NDIMS, int BM, int BN>
 
   // Write the output.
   if (simd_lane_id == 0) {
-    size_t out_column = BN * gid.x + out_offset.x;
+    IdxT out_column = BN * gid.x + out_offset.x;
-    out += full_idx * reduction_stride + out_column;
+    out += full_idx * IdxT(reduction_stride) + out_column;
     if (out_column + n_outputs <= reduction_stride) {
       for (int i = 0; i < n_outputs; i++) {
         out[i] = totals[i];

mlx/backend/metal/kernels/reduction/reduce_row.h

@@ -98,11 +98,11 @@ template <
 METAL_FUNC void per_thread_row_reduce(
     thread U totals[N_WRITES],
     const device T* in,
-    const size_t row_idx,
+    const int64_t row_idx,
     int blocks,
     int extra,
     const constant int* shape,
-    const constant size_t* strides,
+    const constant int64_t* strides,
     const constant int& ndim,
     uint lsize_x,
     uint lid_x) {
@@ -193,18 +193,19 @@ template <
     typename T,
     typename U,
     typename Op,
+    typename IdxT,
     int NDIMS,
     int N_READS = REDUCE_N_READS>
 [[kernel]] void row_reduce_small(
     const device T* in [[buffer(0)]],
     device U* out [[buffer(1)]],
-    const constant size_t& row_size [[buffer(2)]],
+    const constant int64_t& row_size [[buffer(2)]],
-    const constant size_t& non_row_reductions [[buffer(3)]],
+    const constant int64_t& non_row_reductions [[buffer(3)]],
     const constant int* shape [[buffer(4)]],
-    const constant size_t* strides [[buffer(5)]],
+    const constant int64_t* strides [[buffer(5)]],
     const constant int& ndim [[buffer(6)]],
     const constant int* reduce_shape [[buffer(7)]],
-    const constant size_t* reduce_strides [[buffer(8)]],
+    const constant int64_t* reduce_strides [[buffer(8)]],
     const constant int& reduce_ndim [[buffer(9)]],
     uint simd_lane_id [[thread_index_in_simdgroup]],
     uint3 gid [[threadgroup_position_in_grid]],
@@ -214,20 +215,20 @@ template <
   Op op;
 
   U total_val = Op::init;
-  looped_elem_to_loc<NDIMS> loop;
+  LoopedElemToLoc<NDIMS, IdxT, (NDIMS > 2)> loop(reduce_ndim);
 
   // Precompute some row reduction numbers
   const device T* row;
-  int blocks = row_size / N_READS;
+  int blocks = IdxT(row_size) / N_READS;
-  int extra = row_size % N_READS;
+  int extra = IdxT(row_size) % N_READS;
 
   if ((non_row_reductions < 32 && row_size <= 8) || non_row_reductions <= 8) {
     // Simple loop over non_row_reductions and reduce the row in the thread.
-    size_t out_idx = tid.x + tsize.y * size_t(tid.y);
+    IdxT out_idx = tid.x + tsize.y * IdxT(tid.y);
-    in += elem_to_loc(out_idx, shape, strides, ndim);
+    in += elem_to_loc<IdxT>(out_idx, shape, strides, ndim);
 
     for (uint r = 0; r < non_row_reductions; r++) {
-      row = in + loop.location(r, reduce_shape, reduce_strides, reduce_ndim);
+      row = in + loop.location();
       thread_reduce<T, U, Op, N_READS>(total_val, row, blocks, extra);
       loop.next(reduce_shape, reduce_strides);
     }
@@ -236,13 +237,13 @@ template <
   } else {
     // Collaboratively reduce over non_row_reductions in the simdgroup. Each
     // thread reduces every 32nd row and then a simple simd reduce.
-    size_t out_idx = gid.y + gsize.y * size_t(gid.z);
+    IdxT out_idx = gid.y + gsize.y * IdxT(gid.z);
-    in += elem_to_loc(out_idx, shape, strides, ndim);
+    in += elem_to_loc<IdxT>(out_idx, shape, strides, ndim);
 
     loop.next(simd_lane_id, reduce_shape, reduce_strides);
 
     for (uint r = simd_lane_id; r < non_row_reductions; r += simd_size) {
-      row = in + loop.location(r, reduce_shape, reduce_strides, reduce_ndim);
+      row = in + loop.location();
       thread_reduce<T, U, Op, N_READS>(total_val, row, blocks, extra);
       loop.next(simd_size, reduce_shape, reduce_strides);
     }
@@ -259,13 +260,14 @@ template <
     typename T,
     typename U,
     typename Op,
+    typename IdxT = int64_t,
     int N_READS = REDUCE_N_READS,
     int N_WRITES = REDUCE_N_WRITES>
 [[kernel]] void row_reduce_simple(
     const device T* in [[buffer(0)]],
     device U* out [[buffer(1)]],
     const constant size_t& reduction_size [[buffer(2)]],
-    const constant size_t& out_size [[buffer(3)]],
+    const constant int64_t& out_size [[buffer(3)]],
     uint3 gid [[threadgroup_position_in_grid]],
     uint3 gsize [[threadgroups_per_grid]],
     uint3 lid [[thread_position_in_threadgroup]],
@@ -277,15 +279,15 @@ template <
   U totals[N_WRITES];
 
   // Move to the row
-  size_t out_idx = N_WRITES * (gid.y + gsize.y * size_t(gid.z));
+  IdxT out_idx = N_WRITES * (gid.y + gsize.y * IdxT(gid.z));
   if (out_idx + N_WRITES > out_size) {
     out_idx = out_size - N_WRITES;
   }
-  in += out_idx * reduction_size;
+  in += out_idx * IdxT(reduction_size);
   out += out_idx;
 
   // Each thread reduces across the row
-  int blocks = reduction_size / (lsize.x * N_READS);
+  int blocks = IdxT(reduction_size) / (lsize.x * N_READS);
   int extra = reduction_size - blocks * (lsize.x * N_READS);
   per_thread_row_reduce<T, U, Op, N_READS, N_WRITES>(
       totals, in, reduction_size, blocks, extra, lsize.x, lid.x);
@@ -306,18 +308,19 @@ template <
     typename T,
     typename U,
     typename Op,
+    typename IdxT,
     int NDIMS,
     int N_READS = REDUCE_N_READS>
 [[kernel]] void row_reduce_looped(
     const device T* in [[buffer(0)]],
     device U* out [[buffer(1)]],
-    const constant size_t& row_size [[buffer(2)]],
+    const constant int64_t& row_size [[buffer(2)]],
-    const constant size_t& non_row_reductions [[buffer(3)]],
+    const constant int64_t& non_row_reductions [[buffer(3)]],
     const constant int* shape [[buffer(4)]],
-    const constant size_t* strides [[buffer(5)]],
+    const constant int64_t* strides [[buffer(5)]],
     const constant int& ndim [[buffer(6)]],
     const constant int* reduce_shape [[buffer(7)]],
-    const constant size_t* reduce_strides [[buffer(8)]],
+    const constant int64_t* reduce_strides [[buffer(8)]],
     const constant int& reduce_ndim [[buffer(9)]],
     uint3 gid [[threadgroup_position_in_grid]],
     uint3 gsize [[threadgroups_per_grid]],
@@ -330,19 +333,19 @@ template <
   threadgroup U shared_vals[simd_size];
   U total = Op::init;
 
-  size_t out_idx = gid.y + gsize.y * size_t(gid.z);
+  IdxT out_idx = gid.y + gsize.y * IdxT(gid.z);
 
   // lid.x * N_READS breaks the per_thread_row_reduce interface a bit. Maybe it
   // needs a small refactor.
-  in += elem_to_loc(out_idx, shape, strides, ndim) + lid.x * N_READS;
+  in += elem_to_loc<IdxT>(out_idx, shape, strides, ndim) + lid.x * N_READS;
 
-  looped_elem_to_loc<NDIMS> loop;
+  LoopedElemToLoc<NDIMS, IdxT, (NDIMS > 2)> loop(reduce_ndim);
   const device T* row;
-  int blocks = row_size / (lsize.x * N_READS);
+  int blocks = IdxT(row_size) / (lsize.x * N_READS);
   int extra = row_size - blocks * (lsize.x * N_READS);
 
-  for (size_t i = 0; i < non_row_reductions; i++) {
+  for (IdxT i = 0; i < non_row_reductions; i++) {
-    row = in + loop.location(i, reduce_shape, reduce_strides, reduce_ndim);
+    row = in + loop.location();
 
     // Each thread reduces across the row
     U row_total;

mlx/backend/metal/kernels/rms_norm.metal

@@ -3,8 +3,6 @@
 #include <metal_common>
 #include <metal_simdgroup>
 
-#include "mlx/backend/metal/kernels/bf16.h"
-#include "mlx/backend/metal/kernels/defines.h"
 #include "mlx/backend/metal/kernels/utils.h"
 
 using namespace metal;
@@ -17,12 +15,15 @@ template <typename T, int N_READS = RMS_N_READS>
     constant float& eps,
     constant uint& axis_size,
     constant uint& w_stride,
-    threadgroup float* local_inv_mean [[threadgroup(0)]],
-    threadgroup float* local_sums [[threadgroup(1)]],
     uint gid [[threadgroup_position_in_grid]],
     uint lid [[thread_position_in_threadgroup]],
     uint simd_lane_id [[thread_index_in_simdgroup]],
     uint simd_group_id [[simdgroup_index_in_threadgroup]]) {
+  constexpr int SIMD_SIZE = 32;
+
+  threadgroup float local_inv_mean[1];
+  threadgroup float local_sums[SIMD_SIZE];
+
   float acc = 0;
   x += gid * size_t(axis_size) + lid * N_READS;
   w += w_stride * lid * N_READS;
@@ -84,13 +85,15 @@ template <typename T, int N_READS = RMS_N_READS>
     constant float& eps,
     constant uint& axis_size,
     constant uint& w_stride,
-    threadgroup float* local_inv_mean [[threadgroup(0)]],
-    threadgroup float* local_sums [[threadgroup(1)]],
     uint gid [[threadgroup_position_in_grid]],
     uint lid [[thread_position_in_threadgroup]],
     uint lsize [[threads_per_threadgroup]],
     uint simd_lane_id [[thread_index_in_simdgroup]],
     uint simd_group_id [[simdgroup_index_in_threadgroup]]) {
+  constexpr int SIMD_SIZE = 32;
+  threadgroup float local_inv_mean[1];
+  threadgroup float local_sums[SIMD_SIZE];
+
   float acc = 0;
   x += gid * size_t(axis_size) + lid * N_READS;
   w += w_stride * lid * N_READS;
@@ -376,8 +379,6 @@ template <typename T, int N_READS = RMS_N_READS>
       constant float& eps,                                    \
       constant uint& axis_size,                               \
       constant uint& w_stride,                                \
-      threadgroup float* local_inv_mean [[threadgroup(0)]],   \
-      threadgroup float* local_sums [[threadgroup(1)]],       \
       uint gid [[thread_position_in_grid]],                   \
       uint lid [[thread_position_in_threadgroup]],            \
       uint simd_lane_id [[thread_index_in_simdgroup]],        \
@@ -407,8 +408,6 @@ template <typename T, int N_READS = RMS_N_READS>
       constant float& eps,                                       \
       constant uint& axis_size,                                  \
       constant uint& w_stride,                                   \
-      threadgroup float* local_inv_mean [[threadgroup(0)]],      \
-      threadgroup float* local_sums [[threadgroup(1)]],          \
       uint gid [[thread_position_in_grid]],                      \
       uint lid [[thread_position_in_threadgroup]],               \
       uint lsize [[threads_per_threadgroup]],                    \

mlx/backend/metal/kernels/rope.metal

@@ -2,7 +2,6 @@
 
 #include <metal_math>
 
-#include "mlx/backend/metal/kernels/bf16.h"
 #include "mlx/backend/metal/kernels/utils.h"
 template <typename T, bool traditional, bool forward>
 void rope_single_impl(

mlx/backend/metal/kernels/scaled_dot_product_attention.metal

@@ -1,946 +1,16 @@
 #include <metal_stdlib>
 
-#include "mlx/backend/metal/kernels/scaled_dot_product_attention_params.h"
 #include "mlx/backend/metal/kernels/sdpa_vector.h"
-#include "mlx/backend/metal/kernels/steel/defines.h"
-#include "mlx/backend/metal/kernels/steel/gemm/transforms.h"
 #include "mlx/backend/metal/kernels/utils.h"
 
 using namespace metal;
 
-using namespace mlx::steel;
-
-template <
-    typename T,
-    short BROWS,
-    short BCOLS,
-    short dst_ld,
-    short reduction_dim,
-    short tgp_size,
-    short alignment = 1,
-    short n_reads = (BCOLS * BROWS) / (tgp_size),
-    short TCOLS = BCOLS / n_reads,
-    short TROWS = tgp_size / TCOLS>
-struct BlockLoaderFA {
-  STEEL_CONST short n_rows = (BROWS + TROWS - 1) / TROWS;
-  STEEL_CONST short vec_size = n_reads;
-
-  // Leading dimension for src
-  const int src_ld;
-  const int tile_stride;
-
-  // Thread location indices
-  const short thread_idx;
-  const short bi;
-  const short bj;
-
-  // threadgroup and device memory
-  threadgroup T* dst;
-  const device T* src;
-
-  struct alignas(alignment * sizeof(T)) ReadVector {
-    uint8_t v[sizeof(T) * vec_size];
-  };
-
-  /* Constructor */
-  METAL_FUNC BlockLoaderFA(
-      const device T* src_,
-      const int src_ld_,
-      threadgroup T* dst_,
-      ushort simd_group_id [[simdgroup_index_in_threadgroup]],
-      ushort simd_lane_id [[thread_index_in_simdgroup]])
-      : src_ld(src_ld_),
-        tile_stride(reduction_dim ? BCOLS : BROWS * src_ld),
-        thread_idx(simd_group_id * 32 + simd_lane_id),
-        bi(thread_idx / TCOLS),
-        bj(vec_size * (thread_idx % TCOLS)),
-        dst(dst_ + bi * dst_ld + bj),
-        src(src_ + bi * src_ld + bj) {}
-
-  /* Load from device memory into threadgroup memory - without bound checking */
-  METAL_FUNC void load_unsafe() const {
-    STEEL_PRAGMA_UNROLL
-    for (short i = 0; i < BROWS; i += TROWS) {
-      *((threadgroup ReadVector*)(&dst[i * dst_ld])) =
-          *((const device ReadVector*)(&src[i * src_ld]));
-    }
-  }
-
-  /* Load from device memory into threadgroup memory - with bound checking */
-  METAL_FUNC void load_safe(short2 src_tile_dim) const {
-    src_tile_dim = src_tile_dim - short2(bj, bi);
-
-    // Skip loading if thread has no valid reads
-    if (src_tile_dim.x <= 0 || src_tile_dim.y <= 0) {
-      STEEL_PRAGMA_UNROLL
-      for (short i = 0; i < BROWS; i += TROWS) {
-        STEEL_PRAGMA_UNROLL
-        for (short j = 0; j < vec_size; j++) {
-          dst[i * dst_ld + j] = T(0);
-        }
-      }
-      return;
-    }
-
-    // Use fast thread memory for bound checks
-    bool tmp_idx[vec_size];
-    T tmp_val[vec_size];
-
-    STEEL_PRAGMA_UNROLL
-    for (short i = 0; i < BROWS; i += TROWS) {
-      // Make sure tmp_idx only contains valid indices
-      STEEL_PRAGMA_UNROLL
-      for (short j = 0; j < vec_size; j++) {
-        tmp_idx[j] = (i < src_tile_dim.y) && (j < src_tile_dim.x);
-      }
-
-      // Read valid indices into tmp_val
-      STEEL_PRAGMA_UNROLL
-      for (short j = 0; j < vec_size; j++) {
-        tmp_val[j] = src[(tmp_idx[j] ? i * src_ld + j : 0)];
-      }
-
-      // Zero out uneeded values
-      STEEL_PRAGMA_UNROLL
-      for (short j = 0; j < vec_size; j++) {
-        tmp_val[j] = tmp_idx[j] ? tmp_val[j] : T(0);
-      }
-
-      // Copy values to threadgroup memory
-      STEEL_PRAGMA_UNROLL
-      for (short j = 0; j < vec_size; j++) {
-        dst[i * dst_ld + j] = tmp_val[j];
-      }
-    }
-  }
-
-  /* Iteration helper */
-  METAL_FUNC void next() {
-    src += tile_stride;
-  }
-  METAL_FUNC void next(short n) {
-    src += n * tile_stride;
-  }
-};
-
-template <bool M_aligned, bool N_aligned, bool K_aligned>
-struct LoopAlignment {};
-
-template <
-    typename T,
-    typename U,
-    int BM,
-    int BN,
-    int BK,
-    int WM,
-    int WN,
-    bool transpose_a,
-    bool transpose_b,
-    short lda_tgp,
-    short ldb_tgp,
-    typename AccumType = float,
-    typename Epilogue = TransformNone<U, AccumType>>
-struct BlockMMAFA {
-  // Warp tile simdgroup matrix strides along M
-  STEEL_CONST short TM_stride = 8 * WM;
-  // Warp tile simdgroup matrix strides along M
-  STEEL_CONST short TN_stride = 8 * WN;
-
-  // Warp tile size along M
-  STEEL_CONST short TM = BM / TM_stride;
-  // Warp tile size along N
-  STEEL_CONST short TN = BN / TN_stride;
-
-  // Strides of A, B along reduction axis
-  STEEL_CONST short simd_stride_a = {
-      transpose_a ? TM_stride : TM_stride * lda_tgp};
-  STEEL_CONST short simd_stride_b = {
-      transpose_b ? TN_stride * ldb_tgp : TN_stride};
-
-  // Jump between elements
-  STEEL_CONST short jump_a = {transpose_a ? lda_tgp : 1};
-  STEEL_CONST short jump_b = {transpose_b ? ldb_tgp : 1};
-
-  STEEL_CONST short tile_stride_a = {transpose_a ? 8 * lda_tgp : 8};
-  STEEL_CONST short tile_stride_b = {transpose_b ? 8 : 8 * ldb_tgp};
-
-  // Simdgroup matrices
-  simdgroup_matrix<AccumType, 8, 8> Asimd[TM];
-  simdgroup_matrix<AccumType, 8, 8> Bsimd[TN];
-  simdgroup_matrix<AccumType, 8, 8> results[TM * TN] = {
-      simdgroup_matrix<AccumType, 8, 8>(0)};
-
-  // Offsets within threadgroup
-  const short tm;
-  const short tn;
-
-  short sm;
-  short sn;
-
-  ushort sid;
-  ushort slid;
-
-  short As_offset;
-  short Bs_offset;
-
-  /* Constructor */
-  METAL_FUNC BlockMMAFA(
-      ushort simd_group_id [[simdgroup_index_in_threadgroup]],
-      ushort simd_lane_id [[thread_index_in_simdgroup]])
-      : tm(8 * (simd_group_id / WN)), tn(8 * (simd_group_id % WN)) {
-    // Determine thread position in simdgroup matrix
-    short qid = simd_lane_id / 4;
-    slid = simd_lane_id;
-    sid = simd_group_id;
-
-    sm = (qid & 4) + (simd_lane_id / 2) % 4;
-    sn = (qid & 2) * 2 + (simd_lane_id % 2) * 2;
-
-    // Determine thread and simdgroup offset
-    As_offset =
-        transpose_a ? ((sn)*lda_tgp + (tm + sm)) : ((sn) + (tm + sm) * lda_tgp);
-    Bs_offset =
-        transpose_b ? ((tn + sn) * ldb_tgp + (sm)) : ((sm)*ldb_tgp + (tn + sn));
-  }
-
-  /* (BM, BK) X (BK, BN) multiply accumulate function */
-  METAL_FUNC void mma(const threadgroup T* As, const threadgroup T* Bs) {
-    // Adjust for simdgroup and thread location
-    As += As_offset;
-    Bs += Bs_offset;
-
-    // Iterate over BK in blocks of 8
-    STEEL_PRAGMA_UNROLL
-    for (short kk = 0; kk < BK; kk += 8) {
-      simdgroup_barrier(mem_flags::mem_none);
-
-      // Load elements from threadgroup A as simdgroup matrices
-      STEEL_PRAGMA_UNROLL
-      for (short i = 0; i < TM; i++) {
-        Asimd[i].thread_elements()[0] =
-            static_cast<AccumType>(As[i * simd_stride_a + 0]);
-        Asimd[i].thread_elements()[1] =
-            static_cast<AccumType>(As[i * simd_stride_a + jump_a]);
-      }
-
-      simdgroup_barrier(mem_flags::mem_none);
-
-      // Load elements from threadgroup B as simdgroup matrices
-      STEEL_PRAGMA_UNROLL
-      for (short j = 0; j < TN; j++) {
-        Bsimd[j].thread_elements()[0] =
-            static_cast<AccumType>(Bs[j * simd_stride_b + 0]);
-        Bsimd[j].thread_elements()[1] =
-            static_cast<AccumType>(Bs[j * simd_stride_b + jump_b]);
-      }
-
-      simdgroup_barrier(mem_flags::mem_none);
-
-      // Multiply and accumulate into result simdgroup matrices
-      STEEL_PRAGMA_UNROLL
-      for (short i = 0; i < TM; i++) {
-        STEEL_PRAGMA_UNROLL
-        for (short j = 0; j < TN; j++) {
-          short j_serp = (i % 2) ? (TN - 1 - j) : j;
-
-          simdgroup_multiply_accumulate(
-              results[i * TN + j_serp],
-              Asimd[i],
-              Bsimd[j_serp],
-              results[i * TN + j_serp]);
-        }
-      }
-
-      // Progress to next simdgroup tile
-      As += tile_stride_a;
-      Bs += tile_stride_b;
-    }
-  }
-
-  METAL_FUNC void rescale_output(const threadgroup float* Corrections) {
-    // Loop over all simdgroup tiles
-
-    STEEL_PRAGMA_UNROLL
-    for (short i = 0; i < TM; i++) {
-      short row = sm + tm + i * TM_stride;
-      float scale_value = Corrections[row];
-
-      STEEL_PRAGMA_UNROLL
-      for (short j = 0; j < TN; j++) {
-        // Get accumulated result and associated offset in C
-        thread auto& accum = results[i * TN + j].thread_elements();
-        // int offset = (i * TM_stride) * ldc + (j * TN_stride);
-        accum[0] *= scale_value;
-        accum[1] *= scale_value;
-      }
-    }
-  }
-
-  /* Store results from simdgroup_matrix results into device memory */
-  METAL_FUNC void store_result(device U* C, const int ldc) const {
-    // Adjust for simdgroup and thread location
-    C += (sm + tm) * ldc + tn + sn;
-
-    // Loop over all simdgroup tiles
-    STEEL_PRAGMA_UNROLL
-    for (short i = 0; i < TM; i++) {
-      STEEL_PRAGMA_UNROLL
-      for (short j = 0; j < TN; j++) {
-        // Get accumulated result and associated offset in C
-        thread const auto& accum = results[i * TN + j].thread_elements();
-        int offset = (i * TM_stride) * ldc + (j * TN_stride);
-
-        // Apply epilogue
-        U outs[2] = {Epilogue::apply(accum[0]), Epilogue::apply(accum[1])};
-
-        // Write out C
-        C[offset] = outs[0];
-        C[offset + 1] = outs[1];
-      }
-    }
-  }
-
-  METAL_FUNC void store_result_to_tgp_memory(
-      threadgroup U* C,
-      const int ldc,
-      short2 dst_tile_dims) const {
-    // Adjust for simdgroup and thread location
-    C += (sm + tm) * ldc + (tn + sn);
-    dst_tile_dims -= short2(tn + sn, sm + tm);
-
-    STEEL_PRAGMA_UNROLL
-    for (int i = 0; i < TM; i++) {
-      if (i * TM_stride < dst_tile_dims.y) {
-        STEEL_PRAGMA_UNROLL
-        for (int j = 0; j < TN; j++) {
-          // Get accumulated result and associated offset in C
-          thread const auto& accum = results[i * TN + j].thread_elements();
-          int offset = (i * TM_stride) * ldc + (j * TN_stride);
-
-          // Apply epilogue and output C
-          if (j * TN_stride < dst_tile_dims.x) {
-            C[offset] = Epilogue::apply(accum[0]);
-          }
-
-          if (j * TN_stride + 1 < dst_tile_dims.x) {
-            C[offset + 1] = Epilogue::apply(accum[1]);
-          }
-        }
-      }
-    }
-  }
-
-  METAL_FUNC void
-  store_result_safe(device U* C, const int ldc, short2 dst_tile_dims) const {
-    // Adjust for simdgroup and thread location
-    C += (sm + tm) * ldc + (tn + sn);
-    dst_tile_dims -= short2(tn + sn, sm + tm);
-
-    STEEL_PRAGMA_UNROLL
-    for (int i = 0; i < TM; i++) {
-      if (i * TM_stride < dst_tile_dims.y) {
-        STEEL_PRAGMA_UNROLL
-        for (int j = 0; j < TN; j++) {
-          // Get accumulated result and associated offset in C
-          thread const auto& accum = results[i * TN + j].thread_elements();
-          int offset = (i * TM_stride) * ldc + (j * TN_stride);
-
-          // Apply epilogue and output C
-          if (j * TN_stride < dst_tile_dims.x) {
-            C[offset] = Epilogue::apply(accum[0]);
-          }
-
-          if (j * TN_stride + 1 < dst_tile_dims.x) {
-            C[offset + 1] = Epilogue::apply(accum[1]);
-          }
-        }
-      }
-    }
-  }
-
-  /* Store results from simdgroup_matrix results into device memory */
-  METAL_FUNC void store_result(
-      device U* D,
-      const int ldd,
-      const device U* C,
-      const int ldc,
-      const int fdc,
-      thread const Epilogue& epilogue_op) const {
-    // Adjust for simdgroup and thread location
-    C += (sm + tm) * ldc + (tn + sn) * fdc;
-    D += (sm + tm) * ldd + tn + sn;
-
-    // Loop over all simdgroup tiles
-    STEEL_PRAGMA_UNROLL
-    for (short i = 0; i < TM; i++) {
-      STEEL_PRAGMA_UNROLL
-      for (short j = 0; j < TN; j++) {
-        // Get accumulated result and associated offset in C
-        thread const auto& accum = results[i * TN + j].thread_elements();
-        int offset_c = (i * TM_stride) * ldc + (j * TN_stride) * fdc;
-        int offset_d = (i * TM_stride) * ldd + (j * TN_stride);
-
-        // Apply epilogue
-        U outs[2] = {
-            epilogue_op.apply(accum[0], C[offset_c]),
-            epilogue_op.apply(accum[1], C[offset_c + fdc])};
-
-        // Write out D
-        D[offset_d] = outs[0];
-        D[offset_d + 1] = outs[1];
-      }
-    }
-  }
-
-  METAL_FUNC void store_result_safe(
-      device U* D,
-      const int ldd,
-      const device U* C,
-      const int ldc,
-      const int fdc,
-      short2 dst_tile_dims,
-      thread const Epilogue& epilogue_op) const {
-    // Adjust for simdgroup and thread location
-    C += (sm + tm) * ldc + (tn + sn) * fdc;
-    D += (sm + tm) * ldd + tn + sn;
-    dst_tile_dims -= short2(tn + sn, sm + tm);
-
-    STEEL_PRAGMA_UNROLL
-    for (int i = 0; i < TM; i++) {
-      if (i * TM_stride < dst_tile_dims.y) {
-        STEEL_PRAGMA_UNROLL
-        for (int j = 0; j < TN; j++) {
-          // Get accumulated result and associated offset in C
-          thread const auto& accum = results[i * TN + j].thread_elements();
-          int offset_c = (i * TM_stride) * ldc + (j * TN_stride) * fdc;
-          int offset_d = (i * TM_stride) * ldd + (j * TN_stride);
-
-          // Apply epilogue and output C
-          if (j * TN_stride < dst_tile_dims.x) {
-            D[offset_d] = epilogue_op.apply(accum[0], C[offset_c]);
-          }
-
-          if (j * TN_stride + 1 < dst_tile_dims.x) {
-            D[offset_d + 1] = epilogue_op.apply(accum[1], C[offset_c + fdc]);
-          }
-        }
-      }
-    }
-  }
-
-  METAL_FUNC void clear_results() {
-    STEEL_PRAGMA_UNROLL
-    for (int i = 0; i < TM; i++) {
-      STEEL_PRAGMA_UNROLL
-      for (int j = 0; j < TN; j++) {
-        results[i * TN + j] = simdgroup_matrix<AccumType, 8, 8>(0);
-      }
-    }
-  }
-};
-
-template <
-    typename T,
-    typename U,
-    int BM,
-    int BN,
-    int BK,
-    int WM,
-    int WN,
-    bool transpose_q,
-    bool transpose_k,
-    bool transpose_v,
-    bool MN_aligned,
-    bool K_aligned,
-    typename AccumType = typename AccumHelper<T>::accum_type,
-    typename Epilogue = TransformNone<U, AccumType>>
-struct FastAttentionKernel {
-  STEEL_CONST short tgp_padding = 16 / sizeof(T);
-  STEEL_CONST short float_padding = 16 / sizeof(float);
-  STEEL_CONST short tgp_mem_size_q =
-      transpose_q ? BK * (BM + tgp_padding) : BM * (BK + tgp_padding);
-  STEEL_CONST short tgp_mem_size_k =
-      transpose_k ? BK * (BN + tgp_padding) : BN * (BK + tgp_padding);
-  STEEL_CONST short tgp_mem_size_v =
-      transpose_v ? BK * (BN + tgp_padding) : BN * (BK + tgp_padding);
-  STEEL_CONST short tgp_mem_size_s = BM * (BN + tgp_padding);
-
-  // maxes, rowsums, rescale
-  STEEL_CONST short tgp_mem_size_corrections =
-      4 * (BM * sizeof(float) + float_padding);
-
-  STEEL_CONST bool share_kv_smem = transpose_k != transpose_v;
-
-  STEEL_CONST short tgp_mem_size = share_kv_smem
-      ? tgp_mem_size_q + tgp_mem_size_k + tgp_mem_size_s +
-          tgp_mem_size_corrections
-      : tgp_mem_size_q + tgp_mem_size_k + tgp_mem_size_s +
-          tgp_mem_size_corrections + tgp_mem_size_v;
-
-  STEEL_CONST short tgp_size = WM * WN * 32;
-
-  static_assert(transpose_q == false, "Expected Q not transposed.");
-  static_assert(transpose_k == true, "Expected K transposed.");
-  static_assert(transpose_v == false, "Expected V not transposed.");
-  static_assert(tgp_mem_size <= 32768, "Excessive tgp memory requested.");
-
-  using loader_q_t = BlockLoaderFA<
-      T,
-      transpose_q ? BK : BM,
-      transpose_q ? BM : BK,
-      transpose_q ? BM + tgp_padding : BK + tgp_padding,
-      !transpose_q,
-      tgp_size>;
-
-  using loader_k_t = BlockLoaderFA<
-      T,
-      transpose_k ? BN : BK,
-      transpose_k ? BK : BN,
-      transpose_k ? BK + tgp_padding : BN + tgp_padding,
-      transpose_k,
-      tgp_size>;
-
-  using loader_v_t = BlockLoaderFA<
-      T,
-      transpose_v ? BK : BN,
-      transpose_v ? BN : BK,
-      transpose_v ? BN + tgp_padding : BK + tgp_padding,
-      transpose_v,
-      tgp_size>;
-
-  using mma_qk_t = BlockMMAFA<
-      T,
-      U,
-      BM,
-      BN,
-      BK,
-      WM,
-      WN,
-      transpose_q,
-      transpose_k,
-      transpose_q ? BM + tgp_padding : BK + tgp_padding,
-      transpose_k ? BK + tgp_padding : BN + tgp_padding,
-      AccumType,
-      Epilogue>;
-
-  using mma_sv_t = BlockMMAFA<
-      T,
-      U,
-      BM,
-      BK,
-      BN,
-      WM,
-      WN,
-      false,
-      transpose_v,
-      BN + tgp_padding,
-      BK + tgp_padding,
-      AccumType,
-      Epilogue>;
-
-  /* Main kernel function */
-  template <bool M_aligned, bool N_aligned, bool K_aligned_>
-  static METAL_FUNC void gemm_loop(
-      threadgroup T* As [[threadgroup(0)]],
-      threadgroup T* Bs [[threadgroup(1)]],
-      const int gemm_k_iterations,
-      thread loader_k_t& loader_b,
-      thread mma_qk_t& mma_op,
-      thread const short& tgp_bm,
-      thread const short& tgp_bn,
-      LoopAlignment<M_aligned, N_aligned, K_aligned_> l = {}) {
-    // Appease the compiler
-    (void)l;
-    (void)tgp_bm;
-
-    short2 tile_dims_B = transpose_k ? short2(BK, tgp_bn) : short2(tgp_bn, BK);
-
-    // not valid for gemm_k_iterations > 1 (so, BK == d_k)
-    for (int k = 0; k < gemm_k_iterations; k++) {
-      threadgroup_barrier(mem_flags::mem_threadgroup);
-
-      if (N_aligned) {
-        loader_b.load_unsafe();
-      } else {
-        loader_b.load_safe(tile_dims_B);
-      }
-
-      threadgroup_barrier(mem_flags::mem_threadgroup);
-
-      // Multiply and accumulate threadgroup elements
-      mma_op.mma(As, Bs);
-    }
-  }
-
-  static METAL_FUNC void initialize_corrections(
-      threadgroup float* C,
-      uint simd_lane_id,
-      uint simd_group_id) {
-    if (simd_group_id == 0) {
-      threadgroup float* maxes = C;
-      threadgroup float* sums = C + (BM + float_padding);
-      threadgroup float* o_rescale = sums + (BM + float_padding);
-      threadgroup float* output_rescale = o_rescale + (BM + float_padding);
-
-      if (simd_lane_id < BM) {
-        maxes[simd_lane_id] = -INFINITY; // m_i
-        sums[simd_lane_id] = 0.f; // l_i
-        o_rescale[simd_lane_id] = 1.f; // li * exp(mi - mi_new)
-        output_rescale[simd_lane_id] = 1.f; // 1.0 / l_i
-      }
-    }
-  }
-
-  static METAL_FUNC void rescale_ss(
-      threadgroup T* Ss,
-      threadgroup float* Corrections,
-      uint simd_group_id,
-      uint simd_lane_id,
-      short2 local_blocks,
-      float alpha) {
-    if (simd_group_id == 0) {
-      short row_offset = BM + float_padding;
-      threadgroup float* maxes = Corrections;
-      threadgroup float* sums = Corrections + row_offset;
-      threadgroup float* o_rescale = sums + row_offset;
-      threadgroup float* output_scales = o_rescale + row_offset;
-
-      if (simd_lane_id < uint(local_blocks.y)) {
-        float m_i_old = maxes[simd_lane_id];
-        float l_i_old = sums[simd_lane_id];
-
-        float m_i_new = m_i_old;
-        float l_i_new = l_i_old;
-
-        short offset = simd_lane_id * (BN + tgp_padding);
-
-        float m_ij = -INFINITY;
-
-        for (short j = 0; j < local_blocks.x; j++) {
-          float val = alpha * float(Ss[offset + j]);
-          m_ij = max(m_ij, val);
-        }
-
-        m_i_new = max(m_ij, m_i_new);
-
-        float rowsum = 0.f; // lij
-
-        for (short j = 0; j < local_blocks.x; j++) {
-          float val = alpha * float(Ss[offset + j]);
-          float P_i_j = exp(val - m_ij);
-          rowsum += P_i_j;
-          P_i_j = P_i_j * exp(m_ij - m_i_new);
-          Ss[offset + j] = T(P_i_j);
-        }
-
-        l_i_new =
-            exp(m_i_old - m_i_new) * l_i_old + exp(m_ij - m_i_new) * rowsum;
-        maxes[simd_lane_id] = m_i_new;
-        sums[simd_lane_id] = l_i_new;
-        float rescale = l_i_old * exp(m_i_old - m_i_new);
-        o_rescale[simd_lane_id] = rescale;
-        output_scales[simd_lane_id] = 1.0 / l_i_new;
-      }
-    }
-  }
-
-  /* Main kernel function */
-  static METAL_FUNC void run(
-      const device T* Q [[buffer(0)]],
-      const device T* K [[buffer(1)]],
-      const device T* V [[buffer(2)]],
-      device U* O [[buffer(3)]],
-      const constant MLXFastAttentionParams* params [[buffer(4)]],
-      threadgroup T* Qs [[threadgroup(0)]],
-      threadgroup T* Ks [[threadgroup(1)]],
-      threadgroup T* Ss [[threadgroup(2)]],
-      threadgroup T* Vs [[threadgroup(3)]],
-      threadgroup float* Corrections [[threadgroup(4)]],
-      uint simd_lane_id [[thread_index_in_simdgroup]],
-      uint simd_group_id [[simdgroup_index_in_threadgroup]],
-      uint3 tid [[threadgroup_position_in_grid]],
-      uint3 lid [[thread_position_in_threadgroup]]) {
-    // Pacifying compiler
-    (void)lid;
-
-    const int tid_y = ((tid.y) << params->swizzle_log) +
-        ((tid.x) & ((1 << params->swizzle_log) - 1));
-    const int tid_x = (tid.x) >> params->swizzle_log;
-
-    if (params->tiles_n <= tid_x || params->tiles_m <= tid_y) {
-      return;
-    }
-
-    threadgroup_barrier(mem_flags::mem_none);
-
-    // Find block in Q, O; and head in K, V.
-    const int c_row = tid_y * BM;
-
-    Q += transpose_q ? c_row : c_row * params->ldq;
-    thread loader_q_t loader_q(Q, params->ldq, Qs, simd_group_id, simd_lane_id);
-
-    short tgp_bm = min(BM, params->M - c_row);
-    short2 tile_dims_Q = transpose_q ? short2(tgp_bm, BK) : short2(BK, tgp_bm);
-
-    loader_q.load_safe(tile_dims_Q);
-
-    initialize_corrections(Corrections, simd_lane_id, simd_group_id);
-
-    O += c_row * params->ldo;
-
-    // Prepare threadgroup mma operation
-    thread mma_qk_t mma_qk_op(simd_group_id, simd_lane_id);
-    thread mma_sv_t mma_softmax_sv_op(simd_group_id, simd_lane_id);
-    thread loader_k_t loader_k(K, params->ldk, Ks, simd_group_id, simd_lane_id);
-    thread loader_v_t loader_v(V, params->ldv, Vs, simd_group_id, simd_lane_id);
-
-    for (short n_block = 0; n_block < params->gemm_n_iterations_aligned;
-         n_block++) {
-      short c_col = BN;
-
-      // Prepare threadgroup loading operations
-      short gemm_k_iterations = params->gemm_k_iterations_aligned;
-      short tgp_bn_qk = min(BN, params->N - c_col * n_block);
-      threadgroup_barrier(mem_flags::mem_none);
-
-      ///////////////////////////////////////////////////////////////////////////////
-      { // Loop over K - unaligned case
-
-        if (tgp_bm == BM && tgp_bn_qk == BN) {
-          gemm_loop<true, true, K_aligned>(
-              Qs,
-              Ks,
-              gemm_k_iterations,
-              loader_k,
-              mma_qk_op,
-              tgp_bm,
-              tgp_bn_qk);
-        } else if (tgp_bn_qk == BN) {
-          gemm_loop<false, true, K_aligned>(
-              Qs,
-              Ks,
-              gemm_k_iterations,
-              loader_k,
-              mma_qk_op,
-              tgp_bm,
-              tgp_bn_qk);
-
-        } else if (tgp_bm == BM) {
-          gemm_loop<true, false, K_aligned>(
-              Qs,
-              Ks,
-              gemm_k_iterations,
-              loader_k,
-              mma_qk_op,
-              tgp_bm,
-              tgp_bn_qk);
-
-        } else {
-          gemm_loop<false, false, K_aligned>(
-              Qs,
-              Ks,
-              gemm_k_iterations,
-              loader_k,
-              mma_qk_op,
-              tgp_bm,
-              tgp_bn_qk);
-        }
-      }
-
-      mma_qk_op.store_result_to_tgp_memory(
-          Ss, BN + tgp_padding, short2(BN, BM));
-
-      threadgroup_barrier(mem_flags::mem_threadgroup);
-
-      rescale_ss(
-          Ss,
-          Corrections,
-          simd_group_id,
-          simd_lane_id,
-          short2(tgp_bn_qk, tgp_bm),
-          params->alpha);
-
-      loader_v.load_safe(short2(BK, tgp_bn_qk));
-
-      threadgroup_barrier(mem_flags::mem_threadgroup);
-
-      threadgroup float* o_scales = Corrections + 2 * (BM + float_padding);
-      mma_softmax_sv_op.rescale_output(o_scales);
-
-      mma_softmax_sv_op.mma(Ss, Vs);
-
-      threadgroup float* final_output_scales =
-          Corrections + 3 * (BM + float_padding);
-
-      mma_softmax_sv_op.rescale_output(final_output_scales);
-
-      loader_v.next();
-      loader_k.next(BN);
-
-      mma_qk_op.clear_results();
-    }
-
-    threadgroup_barrier(mem_flags::mem_threadgroup);
-    mma_softmax_sv_op.store_result_safe(O, params->ldo, short2(BK, tgp_bm));
-  }
-};
-
-template <
-    typename T,
-    int BM,
-    int BN,
-    int BK,
-    int WM,
-    int WN,
-    bool transpose_q,
-    bool transpose_k,
-    bool transpose_v,
-    bool MN_aligned,
-    bool K_aligned>
-[[kernel, max_total_threads_per_threadgroup(WM* WN * 32)]] void attention(
-    const device T* Q [[buffer(0)]],
-    const device T* K [[buffer(1)]],
-    const device T* V [[buffer(2)]],
-    device T* O [[buffer(3)]],
-    const constant MLXFastAttentionParams* params [[buffer(4)]],
-    const constant int* batch_shape [[buffer(6)]],
-    const constant size_t* batch_strides [[buffer(7)]],
-    uint simd_lane_id [[thread_index_in_simdgroup]],
-    uint simd_group_id [[simdgroup_index_in_threadgroup]],
-    uint3 tid [[threadgroup_position_in_grid]],
-    uint3 lid [[thread_position_in_threadgroup]]) {
-  using attention_kernel = FastAttentionKernel<
-      T,
-      T,
-      BM,
-      BN,
-      BK,
-      WM,
-      WN,
-      transpose_q,
-      transpose_k,
-      transpose_v,
-      MN_aligned,
-      K_aligned>;
-
-  // Adjust for batch
-  if (params->batch_ndim > 1) {
-    const constant size_t* Q_bstrides = batch_strides;
-    const constant size_t* KV_bstrides = batch_strides + params->batch_ndim;
-
-    ulong2 batch_offsets = elem_to_loc_broadcast(
-        tid.z, batch_shape, Q_bstrides, KV_bstrides, params->batch_ndim);
-
-    Q += batch_offsets.x;
-    K += batch_offsets.y;
-    V += batch_offsets.y;
-
-  } else {
-    Q += params->batch_stride_q * tid.z;
-    K += params->batch_stride_k * tid.z;
-    V += params->batch_stride_v * tid.z;
-  }
-
-  // same shape as input
-  O += params->batch_stride_o * tid.z;
-  threadgroup T Qs[attention_kernel::tgp_mem_size_q];
-  threadgroup T Ss[attention_kernel::tgp_mem_size_s];
-  threadgroup float Corrections[attention_kernel::tgp_mem_size_corrections];
-
-  if (attention_kernel::share_kv_smem) {
-    threadgroup T Ks[attention_kernel::tgp_mem_size_k];
-    threadgroup T* Vs = Ks; //[attention_kernel::tgp_mem_size_v];
-    attention_kernel::run(
-        Q,
-        K,
-        V,
-        O,
-        params,
-        Qs,
-        Ks,
-        Ss,
-        Vs,
-        Corrections,
-        simd_lane_id,
-        simd_group_id,
-        tid,
-        lid);
-  } else {
-    threadgroup T Ks[attention_kernel::tgp_mem_size_k];
-    threadgroup T Vs[attention_kernel::tgp_mem_size_v];
-    attention_kernel::run(
-        Q,
-        K,
-        V,
-        O,
-        params,
-        Qs,
-        Ks,
-        Ss,
-        Vs,
-        Corrections,
-        simd_lane_id,
-        simd_group_id,
-        tid,
-        lid);
-  }
-}
-
 // clang-format off
-
-// SDPA full instantiations
-#define instantiate_fast_inference_self_attention_kernel(                   \
-    itype, otype, bm, bn, bk, wm, wn)                                       \
-  template [[host_name("steel_gemm_attention_bm_" #bm "_bn_" #bn "_bk_" #bk \
-                       "_itype_" #itype)]] [[kernel]] void                  \
-  attention<itype, bm, bn, bk, wm, wn, false, true, false, false, true>(    \
-      const device itype* Q [[buffer(0)]],                                  \
-      const device itype* K [[buffer(1)]],                                  \
-      const device itype* V [[buffer(2)]],                                  \
-      device otype* O [[buffer(3)]],                                        \
-      const constant MLXFastAttentionParams* params [[buffer(4)]],          \
-      const constant int* batch_shape [[buffer(6)]],                        \
-      const constant size_t* batch_strides [[buffer(7)]],                   \
-      uint simd_lane_id [[thread_index_in_simdgroup]],                      \
-      uint simd_group_id [[simdgroup_index_in_threadgroup]],                \
-      uint3 tid [[threadgroup_position_in_grid]],                           \
-      uint3 lid [[thread_position_in_threadgroup]]);
-
-instantiate_fast_inference_self_attention_kernel(
-    float,
-    float,
-    16,
-    16,
-    64,
-    2,
-    2);
-instantiate_fast_inference_self_attention_kernel(
-    float,
-    float,
-    16,
-    16,
-    128,
-    2,
-    2);
-instantiate_fast_inference_self_attention_kernel(half, half, 16, 16, 64, 2, 2);
-instantiate_fast_inference_self_attention_kernel(half, half, 16, 16, 128, 2, 2);
-
 // SDPA vector instantiations
-#define instantiate_sdpa_vector(type, head_dim)                              \
+#define instantiate_sdpa_vector(type, head_dim)                                   \
-  template [[host_name("sdpa_vector_" #type "_" #head_dim)]]                 \
+  instantiate_kernel("sdpa_vector_" #type "_" #head_dim, sdpa_vector, type, head_dim)          \
-  [[kernel]] void sdpa_vector<type, head_dim>(                               \
+  instantiate_kernel("sdpa_vector_2pass_1_" #type "_" #head_dim, sdpa_vector_2pass_1, type, head_dim)  \
-      const device type* queries [[buffer(0)]],                              \
+  instantiate_kernel("sdpa_vector_2pass_2_" #type "_" #head_dim, sdpa_vector_2pass_2, type, head_dim)
-      const device type* keys [[buffer(1)]],                                 \
-      const device type* values [[buffer(2)]],                               \
-      device type* out [[buffer(3)]],                                        \
-      const constant int& gqa_factor,                                        \
-      const constant int& N,                                                 \
-      const constant size_t& k_stride,                                       \
-      const constant size_t& v_stride,                                       \
-      const constant float& scale,                                           \
-      uint3 tid [[threadgroup_position_in_grid]],                            \
-      uint simd_gid [[simdgroup_index_in_threadgroup]],                      \
-      uint simd_lid [[thread_index_in_simdgroup]]);
 
 #define instantiate_sdpa_vector_heads(type) \
   instantiate_sdpa_vector(type, 64)         \

mlx/backend/metal/kernels/scaled_dot_product_attention_params.h

@@ -1,42 +0,0 @@
-//
-//  scaled_dot_product_attention_params.h
-//  mlx
-
-#pragma once
-
-struct MLXFastAttentionParams {
-  const int M;
-  const int N;
-  const int K;
-
-  const int ldq; // ldq == ldo
-  const int ldk;
-  const int ldv;
-  const int lds;
-  const int ldo;
-
-  const int tiles_n;
-  const int tiles_m;
-
-  const int batch_stride_q;
-  const int batch_stride_k;
-  const int batch_stride_v;
-  const int batch_stride_o;
-
-  const int swizzle_log;
-  const int gemm_n_iterations_aligned;
-  const int gemm_k_iterations_aligned;
-  const int gemm_sv_m_block_iterations;
-
-  const int batch_ndim;
-  const float alpha;
-};
-
-struct MLXScaledDotProductAttentionParams {
-  // Associated dimensions & transposition information
-  const uint QUERY_SEQUENCE_LENGTH = 1;
-  const uint N_Q_HEADS = 32;
-  const uint N_KV_HEADS = 32;
-  const uint KV_TILES = 1;
-  const float INV_ALPHA = 0.08838834764831843f;
-};

mlx/backend/metal/kernels/scatter.h

@@ -10,16 +10,17 @@ template <
     typename Op,
     int NIDX,
     bool UPD_ROW_CONTIG,
-    int NWORK>
+    int NWORK,
+    typename LocT>
 METAL_FUNC void scatter_impl(
     const device T* updates,
     device mlx_atomic<T>* out,
     const constant int* upd_shape,
-    const constant size_t* upd_strides,
+    const constant int64_t* upd_strides,
     const constant size_t& upd_ndim,
     const constant size_t& upd_size,
     const constant int* out_shape,
-    const constant size_t* out_strides,
+    const constant int64_t* out_strides,
     const constant size_t& out_ndim,
     const constant int* axes,
     const constant size_t& idx_size,
@@ -28,29 +29,30 @@ METAL_FUNC void scatter_impl(
   Op op;
 
   auto ind_idx = gid.y * NWORK;
-  size_t out_offset = 0;
+  LocT out_offset = 0;
   if (upd_size > 1) {
-    out_offset =
+    out_offset = elem_to_loc<LocT>(
-        elem_to_loc(gid.x, upd_shape + indices.ndim, out_strides, out_ndim);
+        gid.x, upd_shape + indices.ndim, out_strides, out_ndim);
   }
 
   for (int j = 0; j < NWORK && ind_idx < idx_size; ++j, ind_idx++) {
-    size_t out_idx = out_offset;
+    LocT out_idx = out_offset;
     for (int i = 0; i < NIDX; ++i) {
       auto idx_loc = indices.row_contiguous[i]
           ? ind_idx
-          : elem_to_loc(
+          : elem_to_loc<LocT>(
                 ind_idx,
                 &indices.shapes[indices.ndim * i],
                 &indices.strides[indices.ndim * i],
                 indices.ndim);
       auto ax = axes[i];
       auto idx_val = offset_neg_idx(indices.buffers[i][idx_loc], out_shape[ax]);
-      out_idx += idx_val * out_strides[ax];
+      out_idx +=
+          static_cast<LocT>(idx_val) * static_cast<LocT>(out_strides[ax]);
     }
-    auto upd_idx = ind_idx * upd_size + gid.x;
+    auto upd_idx = ind_idx * static_cast<LocT>(upd_size) + gid.x;
     if constexpr (!UPD_ROW_CONTIG) {
-      upd_idx = elem_to_loc(upd_idx, upd_shape, upd_strides, upd_ndim);
+      upd_idx = elem_to_loc<LocT>(upd_idx, upd_shape, upd_strides, upd_ndim);
     }
     op.atomic_update(out, updates[upd_idx], out_idx);
   }

mlx/backend/metal/kernels/sdpa_vector.h

@@ -21,8 +21,7 @@ template <typename T, int D>
   constexpr int BN = 32;
   constexpr int BD = 32;
   constexpr int elem_per_thread = D / BD;
-
+  constexpr int stride = BN * D;
-  const int stride = BN * D;
 
   typedef float U;
 
@@ -84,7 +83,6 @@ template <typename T, int D>
     keys += stride;
     values += stride;
   }
-  threadgroup_barrier(mem_flags::mem_threadgroup);
 
   // Each thread has a partial part of the output so we need to combine them.
 
@@ -114,3 +112,181 @@ template <typename T, int D>
     }
   }
 }
+
+template <typename T, int D>
+[[kernel]] void sdpa_vector_2pass_1(
+    const device T* queries [[buffer(0)]],
+    const device T* keys [[buffer(1)]],
+    const device T* values [[buffer(2)]],
+    device float* out [[buffer(3)]],
+    device float* sums [[buffer(4)]],
+    device float* maxs [[buffer(5)]],
+    const constant int& gqa_factor,
+    const constant int& N,
+    const constant size_t& k_stride,
+    const constant size_t& v_stride,
+    const constant float& scale,
+    uint3 tid [[threadgroup_position_in_grid]],
+    uint simd_gid [[simdgroup_index_in_threadgroup]],
+    uint simd_lid [[thread_index_in_simdgroup]]) {
+  constexpr int BN = 8;
+  constexpr int BD = 32;
+  constexpr int elem_per_thread = D / BD;
+  constexpr int stride = BN * D;
+  constexpr int blocks = 32;
+
+  typedef float U;
+
+  thread U q[elem_per_thread];
+  thread U k[elem_per_thread];
+  thread U o[elem_per_thread];
+
+  threadgroup U outputs[BN * BD];
+  threadgroup U max_scores[BN];
+  threadgroup U sum_exp_scores[BN];
+
+  // Adjust positions
+  const int block_idx = tid.z;
+  const int head_idx = tid.y;
+  const int kv_head_idx = head_idx / gqa_factor;
+  queries += head_idx * D + simd_lid * elem_per_thread;
+  keys += kv_head_idx * k_stride + (block_idx * BN + simd_gid) * D +
+      simd_lid * elem_per_thread;
+  values += kv_head_idx * v_stride + (block_idx * BN + simd_gid) * D +
+      simd_lid * elem_per_thread;
+  out += head_idx * blocks * D + block_idx * D + simd_lid * elem_per_thread;
+  sums += head_idx * blocks + block_idx;
+  maxs += head_idx * blocks + block_idx;
+
+  // Read the query and 0 the output accumulator
+  for (int i = 0; i < elem_per_thread; i++) {
+    q[i] = static_cast<U>(scale) * queries[i];
+  }
+  for (int i = 0; i < elem_per_thread; i++) {
+    o[i] = 0;
+  }
+
+  U max_score = -1e9;
+  U sum_exp_score = 0;
+
+  // For each key
+  for (int i = block_idx * BN + simd_gid; i < N; i += blocks * BN) {
+    // Read the key
+    for (int i = 0; i < elem_per_thread; i++) {
+      k[i] = keys[i];
+    }
+
+    // Compute the i-th score
+    U score = 0;
+    for (int i = 0; i < elem_per_thread; i++) {
+      score += q[i] * k[i];
+    }
+    score = simd_sum(score);
+
+    // Update the accumulators
+    U new_max = max(max_score, score);
+    U factor = fast::exp(max_score - new_max);
+    U exp_score = fast::exp(score - new_max);
+
+    max_score = new_max;
+    sum_exp_score = sum_exp_score * factor + exp_score;
+
+    // Update the output accumulator
+    for (int i = 0; i < elem_per_thread; i++) {
+      o[i] = o[i] * factor + exp_score * values[i];
+    }
+
+    // Move the pointers to the next kv
+    keys += blocks * stride;
+    values += blocks * stride;
+  }
+
+  // Each thread has a partial part of the output so we need to combine them.
+
+  // First let's communicate the max and sum_exp
+  if (simd_lid == 0) {
+    max_scores[simd_gid] = max_score;
+    sum_exp_scores[simd_gid] = sum_exp_score;
+  }
+  threadgroup_barrier(mem_flags::mem_threadgroup);
+  max_score = (simd_lid < BN) ? max_scores[simd_lid] : -1e9;
+  U new_max = simd_max(max_score);
+  U factor = fast::exp(max_score - new_max);
+  sum_exp_score = (simd_lid < BN) ? sum_exp_scores[simd_lid] : 0;
+  sum_exp_score = simd_sum(sum_exp_score * factor);
+
+  // Write the sum and new max
+  if (simd_gid == 0) {
+    sums[0] = sum_exp_score;
+    maxs[0] = new_max;
+  }
+
+  // Now we need to aggregate all the outputs
+  for (int i = 0; i < elem_per_thread; i++) {
+    outputs[simd_lid * BN + simd_gid] =
+        o[i] * fast::exp(max_scores[simd_gid] - new_max);
+    threadgroup_barrier(mem_flags::mem_threadgroup);
+
+    // And write the output
+    if (simd_gid == 0) {
+      U output = outputs[simd_lid * BN];
+      for (int j = 1; j < BN; j++) {
+        output += outputs[simd_lid * BN + j];
+      }
+      out[i] = static_cast<T>(output);
+    }
+    threadgroup_barrier(mem_flags::mem_threadgroup);
+  }
+}
+
+template <typename T, int D>
+[[kernel]] void sdpa_vector_2pass_2(
+    const device float* partials [[buffer(0)]],
+    const device float* sums [[buffer(1)]],
+    const device float* maxs [[buffer(2)]],
+    device T* out [[buffer(3)]],
+    uint3 tid [[threadgroup_position_in_grid]],
+    uint simd_gid [[simdgroup_index_in_threadgroup]],
+    uint simd_lid [[thread_index_in_simdgroup]]) {
+  constexpr int BN = 32;
+  constexpr int BD = 32;
+  constexpr int elem_per_thread = D / BD;
+  constexpr int blocks = 32;
+
+  typedef float U;
+
+  thread U o[elem_per_thread];
+  threadgroup U outputs[BN * BD];
+
+  // Adjust positions
+  const int head_idx = tid.y;
+  partials += head_idx * blocks * D + simd_gid * D + simd_lid * elem_per_thread;
+  sums += head_idx * blocks;
+  maxs += head_idx * blocks;
+  out += head_idx * D + simd_gid * elem_per_thread;
+
+  // First everybody reads the max and sum_exp
+  U max_score = maxs[simd_lid];
+  U new_max = simd_max(max_score);
+  U factor = fast::exp(max_score - new_max);
+  U sum_exp_score = simd_sum(sums[simd_lid] * factor);
+
+  // Now read the block into registers and then use shared memory to transpose
+  // it
+  for (int i = 0; i < elem_per_thread; i++) {
+    o[i] = partials[i];
+  }
+  for (int i = 0; i < elem_per_thread; i++) {
+    outputs[simd_lid * BD + simd_gid] = o[i];
+    threadgroup_barrier(mem_flags::mem_threadgroup);
+    o[i] = simd_sum(outputs[simd_gid * BD + simd_lid] * factor) / sum_exp_score;
+    threadgroup_barrier(mem_flags::mem_threadgroup);
+  }
+
+  // And write the output
+  if (simd_lid == 0) {
+    for (int i = 0; i < elem_per_thread; i++) {
+      out[i] = static_cast<T>(o[i]);
+    }
+  }
+}

mlx/backend/metal/kernels/softmax.metal

@@ -6,8 +6,6 @@
 using namespace metal;
 
 // clang-format off
-#include "mlx/backend/metal/kernels/bf16.h"
-#include "mlx/backend/metal/kernels/defines.h"
 #include "mlx/backend/metal/kernels/utils.h"
 #include "mlx/backend/metal/kernels/softmax.h"
 

mlx/backend/metal/kernels/sort.h

@@ -343,8 +343,8 @@ template <
     const constant int& out_stride_sorted_axis [[buffer(4)]],
     const constant int& nc_dim [[buffer(5)]],
     const constant int* nc_shape [[buffer(6)]],
-    const constant size_t* in_nc_strides [[buffer(7)]],
+    const constant int64_t* in_nc_strides [[buffer(7)]],
-    const constant size_t* out_nc_strides [[buffer(8)]],
+    const constant int64_t* out_nc_strides [[buffer(8)]],
     uint3 tid [[threadgroup_position_in_grid]],
     uint3 lid [[thread_position_in_threadgroup]]) {
   using sort_kernel =
@@ -486,7 +486,7 @@ template <
     const constant int& stride_sorted_axis [[buffer(4)]],
     const constant int& nc_dim [[buffer(5)]],
     const constant int* nc_shape [[buffer(6)]],
-    const constant size_t* nc_strides [[buffer(7)]],
+    const constant int64_t* nc_strides [[buffer(7)]],
     uint3 tid [[threadgroup_position_in_grid]],
     uint3 lid [[thread_position_in_threadgroup]]) {
   using sort_kernel = KernelMultiBlockMergeSort<

mlx/backend/metal/kernels/sort.metal

@@ -3,8 +3,6 @@
 #include <metal_stdlib>
 
 // clang-format off
-#include "mlx/backend/metal/kernels/bf16.h"
-#include "mlx/backend/metal/kernels/defines.h"
 #include "mlx/backend/metal/kernels/utils.h"
 #include "mlx/backend/metal/kernels/sort.h"
 

mlx/backend/metal/kernels/steel/attn/attn.h

@@ -0,0 +1,296 @@
+// Copyright © 2024 Apple Inc.
+
+#pragma once
+
+#include "mlx/backend/metal/kernels/steel/attn/loader.h"
+#include "mlx/backend/metal/kernels/steel/attn/mma.h"
+#include "mlx/backend/metal/kernels/steel/attn/params.h"
+#include "mlx/backend/metal/kernels/steel/attn/transforms.h"
+#include "mlx/backend/metal/kernels/steel/gemm/params.h"
+#include "mlx/backend/metal/kernels/steel/utils.h"
+
+using namespace metal;
+
+///////////////////////////////////////////////////////////////////////////////
+// GEMM kernel class
+///////////////////////////////////////////////////////////////////////////////
+
+namespace mlx {
+namespace steel {
+
+template <bool M_aligned, bool N_aligned, bool K_aligned>
+struct LoopAlignment {};
+
+template <
+    typename T,
+    typename U,
+    int BM,
+    int BN,
+    int BK,
+    int WM,
+    int WN,
+    bool transpose_a,
+    bool transpose_b,
+    bool MN_aligned,
+    bool K_aligned,
+    typename AccumType = typename AccumHelper<T>::accum_type,
+    typename Epilogue = TransformNone<U, AccumType>>
+struct GEMMKernel {
+  STEEL_CONST short tgp_padding_a = 16 / sizeof(T);
+  STEEL_CONST short tgp_padding_b = 16 / sizeof(T);
+  STEEL_CONST short tgp_mem_size_a =
+      transpose_a ? BK * (BM + tgp_padding_a) : BM * (BK + tgp_padding_a);
+  STEEL_CONST short tgp_mem_size_b =
+      transpose_b ? BN * (BK + tgp_padding_b) : BK * (BN + tgp_padding_b);
+  STEEL_CONST short tgp_mem_size = tgp_mem_size_a + tgp_mem_size_b;
+
+  STEEL_CONST short tgp_size = WM * WN * 32;
+
+  using loader_a_t = BlockLoader<
+      T,
+      transpose_a ? BK : BM,
+      transpose_a ? BM : BK,
+      transpose_a ? BM + tgp_padding_a : BK + tgp_padding_a,
+      !transpose_a,
+      tgp_size>;
+  using loader_b_t = BlockLoader<
+      T,
+      transpose_b ? BN : BK,
+      transpose_b ? BK : BN,
+      transpose_b ? BK + tgp_padding_b : BN + tgp_padding_b,
+      transpose_b,
+      tgp_size>;
+  using mma_t = BlockMMA<
+      T,
+      U,
+      BM,
+      BN,
+      BK,
+      WM,
+      WN,
+      transpose_a,
+      transpose_b,
+      transpose_a ? BM + tgp_padding_a : BK + tgp_padding_a,
+      transpose_b ? BK + tgp_padding_b : BN + tgp_padding_b,
+      AccumType,
+      Epilogue>;
+
+  /* Main kernel function */
+  template <bool M_aligned, bool N_aligned, bool K_aligned_>
+  static METAL_FUNC void gemm_loop(
+      threadgroup T* As [[threadgroup(0)]],
+      threadgroup T* Bs [[threadgroup(1)]],
+      const int gemm_k_iterations,
+      thread loader_a_t& loader_a,
+      thread loader_b_t& loader_b,
+      thread mma_t& mma_op,
+      thread const short& tgp_bm,
+      thread const short& tgp_bn,
+      thread const short& lbk,
+      LoopAlignment<M_aligned, N_aligned, K_aligned_> l = {}) {
+    // Appease the compiler
+    (void)l;
+
+    short2 tile_dims_A = transpose_a ? short2(tgp_bm, BK) : short2(BK, tgp_bm);
+
+    short2 tile_dims_B = transpose_b ? short2(BK, tgp_bn) : short2(tgp_bn, BK);
+
+    for (int k = 0; k < gemm_k_iterations; k++) {
+      threadgroup_barrier(mem_flags::mem_threadgroup);
+      // Load elements into threadgroup
+      if (M_aligned) {
+        loader_a.load_unsafe();
+      } else {
+        loader_a.load_safe(tile_dims_A);
+      }
+
+      if (N_aligned) {
+        loader_b.load_unsafe();
+      } else {
+        loader_b.load_safe(tile_dims_B);
+      }
+
+      threadgroup_barrier(mem_flags::mem_threadgroup);
+
+      // Multiply and accumulate threadgroup elements
+      mma_op.mma(As, Bs);
+
+      // Prepare for next iteration
+      loader_a.next();
+      loader_b.next();
+    }
+
+    if (!K_aligned_) {
+      threadgroup_barrier(mem_flags::mem_threadgroup);
+
+      short2 tile_dims_A_last =
+          transpose_a ? short2(tgp_bm, lbk) : short2(lbk, tgp_bm);
+      short2 tile_dims_B_last =
+          transpose_b ? short2(lbk, tgp_bn) : short2(tgp_bn, lbk);
+
+      loader_a.load_safe(tile_dims_A_last);
+      loader_b.load_safe(tile_dims_B_last);
+
+      threadgroup_barrier(mem_flags::mem_threadgroup);
+
+      mma_op.mma(As, Bs);
+    }
+  }
+
+  /* Main kernel function */
+  static METAL_FUNC void run(
+      const device T* A [[buffer(0)]],
+      const device T* B [[buffer(1)]],
+      device U* D [[buffer(2)]],
+      const constant GEMMParams* params [[buffer(3)]],
+      threadgroup T* As [[threadgroup(0)]],
+      threadgroup T* Bs [[threadgroup(1)]],
+      uint simd_lane_id [[thread_index_in_simdgroup]],
+      uint simd_group_id [[simdgroup_index_in_threadgroup]],
+      uint3 tid [[threadgroup_position_in_grid]],
+      uint3 lid [[thread_position_in_threadgroup]]) {
+    // Pacifying compiler
+    (void)lid;
+
+    const int tid_y = ((tid.y) << params->swizzle_log) +
+        ((tid.x) & ((1 << params->swizzle_log) - 1));
+    const int tid_x = (tid.x) >> params->swizzle_log;
+
+    if (params->tiles_n <= tid_x || params->tiles_m <= tid_y) {
+      return;
+    }
+
+    threadgroup_barrier(mem_flags::mem_none);
+
+    // Find block in A, B, C
+    const int c_row = tid_y * BM;
+    const int c_col = tid_x * BN;
+    const size_t c_row_long = size_t(c_row);
+    const size_t c_col_long = size_t(c_col);
+
+    A += transpose_a ? c_row_long : c_row_long * params->lda;
+    B += transpose_b ? c_col_long * params->ldb : c_col_long;
+    D += c_row_long * params->ldd + c_col_long;
+
+    // Prepare threadgroup loading operations
+    thread loader_a_t loader_a(A, params->lda, As, simd_group_id, simd_lane_id);
+    thread loader_b_t loader_b(B, params->ldb, Bs, simd_group_id, simd_lane_id);
+
+    // Prepare threadgroup mma operation
+    thread mma_t mma_op(simd_group_id, simd_lane_id);
+
+    int gemm_k_iterations = params->gemm_k_iterations_aligned;
+
+    ///////////////////////////////////////////////////////////////////////////////
+    // MNK aligned loop
+    if (MN_aligned) {
+      for (int k = 0; k < gemm_k_iterations; k++) {
+        threadgroup_barrier(mem_flags::mem_threadgroup);
+        // Load elements into threadgroup
+        loader_a.load_unsafe();
+        loader_b.load_unsafe();
+
+        threadgroup_barrier(mem_flags::mem_threadgroup);
+
+        // Multiply and accumulate threadgroup elements
+        mma_op.mma(As, Bs);
+
+        // Prepare for next iteration
+        loader_a.next();
+        loader_b.next();
+      }
+
+      threadgroup_barrier(mem_flags::mem_none);
+
+      // Loop tail
+      if (!K_aligned) {
+        int lbk = params->K - params->gemm_k_iterations_aligned * BK;
+        short2 tile_dims_A = transpose_a ? short2(BM, lbk) : short2(lbk, BM);
+        short2 tile_dims_B = transpose_b ? short2(lbk, BN) : short2(BN, lbk);
+
+        loader_a.load_safe(tile_dims_A);
+        loader_b.load_safe(tile_dims_B);
+
+        threadgroup_barrier(mem_flags::mem_threadgroup);
+
+        mma_op.mma(As, Bs);
+      }
+
+      // Store results to device memory
+      mma_op.store_result(D, params->ldd);
+      return;
+
+    }
+    ///////////////////////////////////////////////////////////////////////////////
+    // MN unaligned loop
+    else { // Loop over K - unaligned case
+      short tgp_bm = min(BM, params->M - c_row);
+      short tgp_bn = min(BN, params->N - c_col);
+      short leftover_bk = params->K - params->gemm_k_iterations_aligned * BK;
+
+      if (tgp_bm == BM && tgp_bn == BN) {
+        gemm_loop<true, true, K_aligned>(
+            As,
+            Bs,
+            gemm_k_iterations,
+            loader_a,
+            loader_b,
+            mma_op,
+            tgp_bm,
+            tgp_bn,
+            leftover_bk);
+
+        mma_op.store_result(D, params->ldd);
+        return;
+
+      } else if (tgp_bn == BN) {
+        gemm_loop<false, true, K_aligned>(
+            As,
+            Bs,
+            gemm_k_iterations,
+            loader_a,
+            loader_b,
+            mma_op,
+            tgp_bm,
+            tgp_bn,
+            leftover_bk);
+
+        mma_op.store_result_safe(D, params->ldd, short2(tgp_bn, tgp_bm));
+        return;
+
+      } else if (tgp_bm == BM) {
+        gemm_loop<true, false, K_aligned>(
+            As,
+            Bs,
+            gemm_k_iterations,
+            loader_a,
+            loader_b,
+            mma_op,
+            tgp_bm,
+            tgp_bn,
+            leftover_bk);
+
+        mma_op.store_result_safe(D, params->ldd, short2(tgp_bn, tgp_bm));
+        return;
+
+      } else {
+        gemm_loop<false, false, K_aligned>(
+            As,
+            Bs,
+            gemm_k_iterations,
+            loader_a,
+            loader_b,
+            mma_op,
+            tgp_bm,
+            tgp_bn,
+            leftover_bk);
+
+        mma_op.store_result_safe(D, params->ldd, short2(tgp_bn, tgp_bm));
+        return;
+      }
+    }
+  }
+};
+
+} // namespace steel
+} // namespace mlx

mlx/backend/metal/kernels/steel/attn/kernels/steel_attention.h

@@ -0,0 +1,349 @@
+// Copyright © 2024 Apple Inc.
+
+using namespace mlx::steel;
+
+///////////////////////////////////////////////////////////////////////////////
+// GEMM kernels
+///////////////////////////////////////////////////////////////////////////////
+
+constant bool align_Q [[function_constant(200)]];
+constant bool align_K [[function_constant(201)]];
+
+template <typename T>
+struct TransformScale {
+  T scale;
+  METAL_FUNC TransformScale(T scale_) : scale(scale_) {}
+
+  METAL_FUNC T apply(T x) const {
+    return scale * x;
+  }
+};
+
+struct MaxOp {
+  template <typename T>
+  METAL_FUNC static constexpr T apply(T x, T y) {
+    return metal::max(x, y);
+  }
+};
+
+struct SumOp {
+  template <typename T>
+  METAL_FUNC static constexpr T apply(T x, T y) {
+    return x + y;
+  }
+};
+
+struct MulOp {
+  template <typename T>
+  METAL_FUNC static constexpr T apply(T x, T y) {
+    return x * y;
+  }
+};
+
+struct SubOp {
+  template <typename T>
+  METAL_FUNC static constexpr T apply(T x, T y) {
+    return x - y;
+  }
+};
+
+struct ExpSubOp {
+  template <typename T>
+  METAL_FUNC static constexpr T apply(T x, T y) {
+    return fast::exp(x - y);
+  }
+};
+
+struct DivOp {
+  template <typename T>
+  METAL_FUNC static constexpr T apply(T x, T y) {
+    return x / y;
+  }
+};
+
+// clang-format off
+template <
+    typename T,
+    int BQ,
+    int BK,
+    int BD,
+    int WM,
+    int WN,
+    typename AccumType = float>
+[[kernel, max_total_threads_per_threadgroup(WM * WN * 32)]] void attention(
+    const device T* Q [[buffer(0)]],
+    const device T* K [[buffer(1)]],
+    const device T* V [[buffer(2)]],
+    device T* O [[buffer(3)]],
+    const constant AttnParams* params [[buffer(4)]],
+    uint simd_lane_id [[thread_index_in_simdgroup]],
+    uint simd_group_id [[simdgroup_index_in_threadgroup]],
+    uint3 tid [[threadgroup_position_in_grid]],
+    uint3 lid [[thread_position_in_threadgroup]]) { // clang-format on
+
+  // Pacifying compiler
+  (void)lid;
+
+  // Move to correct block
+  ulong3 tidl{tid.x, tid.y, tid.z};
+
+  Q += tidl.z * params->Q_strides[0] + // Batch
+      tidl.y * params->Q_strides[1] + // Head
+      tidl.x * BQ * params->Q_strides[2]; // Seqeunce
+
+  ulong kv_head_idx = int(tid.y) / params->gqa_factor;
+  K += tidl.z * params->K_strides[0] + // Batch
+      kv_head_idx * params->K_strides[1]; // Head
+
+  V += tidl.z * params->V_strides[0] + // Batch
+      kv_head_idx * params->V_strides[1]; // Head
+
+  O += tidl.z * params->O_strides[0] + // Batch
+      tidl.y * params->O_strides[1] + // Head
+      tidl.x * BQ * params->O_strides[2]; // Seqeunce
+
+  // Prepare threadgroup memory
+  constexpr short padQ = 0; // 16 / sizeof(T);
+  constexpr short padK = 0; // 16 / sizeof(T);
+  constexpr short padV = 0; // 16 / sizeof(T);
+
+  constexpr short LDQ_tgp = BD + padQ;
+  constexpr short LDK_tgp = BK + padK;
+  constexpr short LDV_tgp = BD + padV;
+
+  threadgroup T Qs[BQ * (BD + padQ)];
+  threadgroup T Ks[(BK + padK) * BD];
+  threadgroup T Vs[BK * (BD + padV)];
+
+  // Prepare block loaders
+  using QBlockLoader = BlockLoaderT<
+      /* typename T = */ T,
+      /* short BROWS = */ BQ,
+      /* short BCOLS = */ BD,
+      /* short kDstStrRow = */ LDQ_tgp,
+      /* short kDstStrCol = */ 1,
+      /* short reduction_dim = */ 1,
+      /* short tgp_size = */ WM * WN * 32>;
+
+  // K is loaded in transposed
+  using KBlockLoader = BlockLoaderT<
+      /* typename T = */ T,
+      /* short BROWS = */ BK,
+      /* short BCOLS = */ BD,
+      /* short kDstStrRow = */ 1,
+      /* short kDstStrCol = */ LDK_tgp,
+      /* short reduction_dim = */ 0,
+      /* short tgp_size = */ WM * WN * 32>;
+
+  using VBlockLoader = BlockLoaderT<
+      /* typename T = */ T,
+      /* short BROWS = */ BK,
+      /* short BCOLS = */ BD,
+      /* short kDstStrRow = */ LDV_tgp,
+      /* short kDstStrCol = */ 1,
+      /* short reduction_dim = */ 0,
+      /* short tgp_size = */ WM * WN * 32>;
+
+  QBlockLoader loader_q(
+      Q, params->Q_strides[2], Qs, simd_group_id, simd_lane_id);
+  KBlockLoader loader_k(
+      K, params->K_strides[2], Ks, simd_group_id, simd_lane_id);
+  VBlockLoader loader_v(
+      V, params->V_strides[2], Vs, simd_group_id, simd_lane_id);
+
+  TransformScale<T> ts(static_cast<T>(params->scale));
+
+  // Prepare MMA tiles
+  constexpr short kFragSize = 8; // MMAFrag size
+  using MMAFrag_acc_t = BaseMMAFrag<AccumType, kFragSize, kFragSize>;
+
+  constexpr int kNWarps = WM * WN;
+  static_assert(
+      BQ >= (kNWarps * kFragSize) && BQ % (kNWarps * kFragSize) == 0,
+      "Each simdgroup must host atleast 1 simdgroup matrix along Q sequence.");
+
+  // Q seq frags per warp
+  constexpr int TQ = BQ / (kNWarps * kFragSize);
+  // KV sequence frags (all warps load the same frags)
+  constexpr int TK = BK / kFragSize;
+  // HeadDim frags (all warps load the same frags)
+  constexpr int TD = BD / kFragSize;
+
+  static_assert(TQ == 1, "Check TQ");
+
+  MMATile<AccumType, TQ, 1, MMAFrag_acc_t> Qtile;
+  MMATile<AccumType, 1, TK, MMAFrag_acc_t> Ktile;
+  MMATile<AccumType, TQ, TK, MMAFrag_acc_t> Stile;
+  MMATile<AccumType, TK, TD, MMAFrag_acc_t> Vtile;
+  MMATile<AccumType, TQ, TD, MMAFrag_acc_t> Otile;
+
+  Otile.clear();
+
+  // Prepare mma tile offsets
+  const short2 simd_coord = MMAFrag_acc_t::get_coord(simd_lane_id);
+  const short sm = simd_coord.y;
+  const short sn = simd_coord.x;
+  const short tm = kFragSize * TQ * simd_group_id;
+
+  const short Qs_offset = (tm + sm) * LDQ_tgp + sn;
+  const short Ks_offset = sm * LDK_tgp + sn;
+  const short Vs_offset = sm * LDV_tgp + sn;
+
+  constexpr short Qs_tile_stride = kFragSize;
+  constexpr short Ks_tile_stride = kFragSize * LDK_tgp;
+
+  threadgroup_barrier(mem_flags::mem_threadgroup);
+
+  // Load Q blocks apply scale
+  if (!align_Q && int(tid.x) == (params->NQ_aligned)) {
+    loader_q.load_safe(short2(BD, params->qL - params->NQ_aligned * BQ));
+  } else {
+    loader_q.load_unsafe();
+  }
+  loader_q.apply_inplace_op(ts);
+
+  // Init row reduction variables
+  constexpr short kRowsPT = decltype(Stile)::kRowsPerThread;
+
+  AccumType max_score[kRowsPT];
+  AccumType sum_score[kRowsPT] = {0};
+
+  // Init to -Inf
+  STEEL_PRAGMA_UNROLL
+  for (short i = 0; i < kRowsPT; ++i) {
+    max_score[i] = Limits<AccumType>::min;
+  }
+
+  // Loop over KV seq length
+  for (int kb = 0; kb < params->NK; kb++) {
+    // Load K block and apply scale
+    threadgroup_barrier(mem_flags::mem_threadgroup);
+    if (!align_K && kb == (params->NK_aligned)) {
+      loader_k.load_safe(short2(BD, params->kL - params->NK_aligned * BK));
+    } else {
+      loader_k.load_unsafe();
+    }
+
+    threadgroup_barrier(mem_flags::mem_threadgroup);
+
+    // Do S = Q @ K.T
+    Stile.clear();
+
+    for (short dd = 0; dd < TD; dd++) {
+      simdgroup_barrier(mem_flags::mem_none);
+
+      Qtile.template load<T, 1, 1, LDQ_tgp, 1>(
+          &Qs[Qs_offset + dd * Qs_tile_stride]);
+      Ktile.template load<T, 1, 1, LDK_tgp, 1>(
+          &Ks[Ks_offset + dd * Ks_tile_stride]);
+
+      simdgroup_barrier(mem_flags::mem_none);
+
+      tile_matmad(Stile, Qtile, Ktile, Stile);
+    }
+
+    // Mask out of length sequence
+    if (!align_K && kb == (params->NK_aligned)) {
+      using stile_t = decltype(Stile);
+      using selem_t = typename stile_t::elem_type;
+      constexpr auto neg_inf = -metal::numeric_limits<selem_t>::infinity();
+      const short lim = params->kL - params->NK_aligned * BK;
+
+      STEEL_PRAGMA_UNROLL
+      for (short i = 0; i < stile_t::kTileRows; i++) {
+        STEEL_PRAGMA_UNROLL
+        for (short j = 0; j < stile_t::kTileCols; j++) {
+          short col_pos = sn + (j * stile_t::kFragCols);
+          STEEL_PRAGMA_UNROLL
+          for (short jj = 0; jj < stile_t::MMAFrag_t::kElemCols; jj++) {
+            if ((col_pos + jj) >= lim) {
+              Stile.frag_at(i, j)[jj] = neg_inf;
+            }
+          }
+        }
+      }
+    }
+
+    simdgroup_barrier(mem_flags::mem_none);
+
+    // Load V blocks
+    if (!align_K && kb == (params->NK_aligned)) {
+      loader_v.load_safe(short2(BD, params->kL - params->NK_aligned * BK));
+    } else {
+      loader_v.load_unsafe();
+    }
+
+    // Do softmax
+
+    // Temp variables
+    AccumType new_max[kRowsPT];
+    AccumType factor[kRowsPT];
+    STEEL_PRAGMA_UNROLL
+    for (short i = 0; i < kRowsPT; ++i) {
+      new_max[i] = max_score[i];
+    }
+
+    // Row max
+    Stile.template row_reduce<MaxOp>(new_max);
+
+    // exp(Si - rowmax(Si))
+    Stile.template row_bin_op<ExpSubOp>(new_max);
+
+    // Factor exp(rowmax(Si) - rowmax(Si-1))
+    STEEL_PRAGMA_UNROLL
+    for (short i = 0; i < kRowsPT; ++i) {
+      factor[i] = fast::exp(max_score[i] - new_max[i]);
+    }
+
+    // Save max for next iteration
+    STEEL_PRAGMA_UNROLL
+    for (short i = 0; i < kRowsPT; ++i) {
+      max_score[i] = new_max[i];
+    }
+
+    // Row Sum
+    AccumType sum_score_tmp[kRowsPT] = {0};
+    Stile.template row_reduce<SumOp>(sum_score_tmp);
+
+    // Update norm
+    STEEL_PRAGMA_UNROLL
+    for (short i = 0; i < kRowsPT; ++i) {
+      sum_score[i] = sum_score[i] * factor[i] + sum_score_tmp[i];
+    }
+
+    // Update O
+    Otile.template row_bin_op<MulOp>(factor);
+
+    // Load V into registers
+    threadgroup_barrier(mem_flags::mem_threadgroup);
+    Vtile.template load<T, 1, 1, LDV_tgp, 1>(&Vs[Vs_offset]);
+
+    simdgroup_barrier(mem_flags::mem_none);
+
+    // Do O = S @ V
+    tile_matmad(Otile, Stile, Vtile, Otile);
+
+    // Prepare for next iteration
+    loader_k.next();
+    loader_v.next();
+  }
+
+  // Normalize output
+  Otile.template row_bin_op<DivOp>(sum_score);
+  threadgroup_barrier(mem_flags::mem_none);
+
+  // Store results
+  O += (tm + sm) * params->O_strides[2] + sn;
+
+  if (!align_Q && int(tid.x) == (params->NQ_aligned)) {
+    auto dst_tile_dims =
+        short2(BD - sn, params->qL - BQ * params->NQ_aligned - (tm + sm));
+
+    if (dst_tile_dims.x <= 0 || dst_tile_dims.y <= 0)
+      return;
+
+    Otile.template store_safe<T, 1, 1>(O, params->O_strides[2], dst_tile_dims);
+  } else {
+    Otile.template store<T, 1, 1>(O, params->O_strides[2]);
+  }
+}

mlx/backend/metal/kernels/steel/attn/kernels/steel_attention.metal

@@ -0,0 +1,31 @@
+// Copyright © 2024 Apple Inc.
+
+// clang-format off
+#include "mlx/backend/metal/kernels/utils.h"
+
+#include "mlx/backend/metal/kernels/steel/attn/attn.h"
+#include "mlx/backend/metal/kernels/steel/attn/kernels/steel_attention.h"
+
+#define instantiate_attn(tname, dtype, bq, bk, bd, wm, wn) \
+  template [[host_name("steel_attention_" #tname "_bq" #bq "_bk" #bk "_bd" #bd "_wm" #wm "_wn" #wn)]] \
+  [[kernel]] void attention<dtype, bq, bk, bd, wm, wn, float>( \
+      const device dtype* Q [[buffer(0)]], \
+      const device dtype* K [[buffer(1)]], \
+      const device dtype* V [[buffer(2)]], \
+      device dtype* O [[buffer(3)]],\
+      const constant AttnParams* params [[buffer(4)]], \
+      uint simd_lane_id [[thread_index_in_simdgroup]], \
+      uint simd_group_id [[simdgroup_index_in_threadgroup]], \
+      uint3 tid [[threadgroup_position_in_grid]], \
+      uint3 lid [[thread_position_in_threadgroup]]);
+
+#define instantiate_attn_shapes_helper(iname, itype) \
+    instantiate_attn(iname, itype, 32, 16, 128, 4, 1) \
+    instantiate_attn(iname, itype, 32, 32,  80, 4, 1) \
+    instantiate_attn(iname, itype, 32, 32,  64, 4, 1)
+
+instantiate_attn_shapes_helper(float16, half);
+instantiate_attn_shapes_helper(bfloat16, bfloat16_t);
+
+instantiate_attn_shapes_helper(float32, float);
+// clang-format on

mlx/backend/metal/kernels/steel/attn/loader.h

@@ -0,0 +1,264 @@
+// Copyright © 2024 Apple Inc.
+
+#pragma once
+
+#include "mlx/backend/metal/kernels/steel/defines.h"
+
+///////////////////////////////////////////////////////////////////////////////
+// Loading helper
+///////////////////////////////////////////////////////////////////////////////
+
+namespace mlx {
+namespace steel {
+
+template <
+    typename T,
+    short BROWS,
+    short BCOLS,
+    short dst_ld,
+    short reduction_dim,
+    short tgp_size,
+    short alignment = 1,
+    short n_reads = (BCOLS * BROWS) / (tgp_size),
+    short TCOLS = BCOLS / n_reads,
+    short TROWS = tgp_size / TCOLS>
+struct BlockLoader {
+  STEEL_CONST short n_rows = (BROWS + TROWS - 1) / TROWS;
+  STEEL_CONST short vec_size = n_reads;
+
+  // Leading dimension for src
+  const int src_ld;
+  const int tile_stride;
+
+  // Thread location indices
+  const short thread_idx;
+  const short bi;
+  const short bj;
+
+  // threadgroup and device memory
+  threadgroup T* dst;
+  const device T* src;
+
+  struct alignas(alignment * sizeof(T)) ReadVector {
+    uint8_t v[sizeof(T) * vec_size];
+  };
+
+  /* Constructor */
+  METAL_FUNC BlockLoader(
+      const device T* src_,
+      const int src_ld_,
+      threadgroup T* dst_,
+      ushort simd_group_id [[simdgroup_index_in_threadgroup]],
+      ushort simd_lane_id [[thread_index_in_simdgroup]])
+      : src_ld(src_ld_),
+        tile_stride(reduction_dim ? BCOLS : BROWS * src_ld),
+        thread_idx(simd_group_id * 32 + simd_lane_id),
+        bi(thread_idx / TCOLS),
+        bj(vec_size * (thread_idx % TCOLS)),
+        dst(dst_ + bi * dst_ld + bj),
+        src(src_ + bi * src_ld + bj) {}
+
+  /* Apply operation to threadgroup without bound checking */
+  template <typename UnaryOp>
+  METAL_FUNC void apply_inplace_op(thread const UnaryOp& op) const {
+    STEEL_PRAGMA_UNROLL
+    for (short i = 0; i < BROWS; i += TROWS) {
+      STEEL_PRAGMA_UNROLL
+      for (short j = 0; j < vec_size; j++) {
+        dst[i * dst_ld + j] = op.apply(dst[i * dst_ld + j]);
+      }
+    }
+  }
+
+  /* Load from device memory into threadgroup memory - without bound checking */
+  METAL_FUNC void load_unsafe() const {
+    STEEL_PRAGMA_UNROLL
+    for (short i = 0; i < BROWS; i += TROWS) {
+      *((threadgroup ReadVector*)(&dst[i * dst_ld])) =
+          *((const device ReadVector*)(&src[i * src_ld]));
+    }
+  }
+
+  /* Load from device memory into threadgroup memory - with bound checking */
+  METAL_FUNC void load_safe(short2 src_tile_dim) const {
+    src_tile_dim = src_tile_dim - short2(bj, bi);
+
+    // Skip loading if thread has no valid reads
+    if (src_tile_dim.x <= 0 || src_tile_dim.y <= 0) {
+      STEEL_PRAGMA_UNROLL
+      for (short i = 0; i < BROWS; i += TROWS) {
+        STEEL_PRAGMA_UNROLL
+        for (short j = 0; j < vec_size; j++) {
+          dst[i * dst_ld + j] = T(0);
+        }
+      }
+      return;
+    }
+
+    // Use fast thread memory for bound checks
+    bool tmp_idx[vec_size];
+    T tmp_val[vec_size];
+
+    STEEL_PRAGMA_UNROLL
+    for (short i = 0; i < BROWS; i += TROWS) {
+      // Make sure tmp_idx only contains valid indices
+      STEEL_PRAGMA_UNROLL
+      for (short j = 0; j < vec_size; j++) {
+        tmp_idx[j] = (i < src_tile_dim.y) && (j < src_tile_dim.x);
+      }
+
+      // Read valid indices into tmp_val
+      STEEL_PRAGMA_UNROLL
+      for (short j = 0; j < vec_size; j++) {
+        tmp_val[j] = src[(tmp_idx[j] ? i * src_ld + j : 0)];
+      }
+
+      // Zero out uneeded values
+      STEEL_PRAGMA_UNROLL
+      for (short j = 0; j < vec_size; j++) {
+        tmp_val[j] = tmp_idx[j] ? tmp_val[j] : T(0);
+      }
+
+      // Copy values to threadgroup memory
+      STEEL_PRAGMA_UNROLL
+      for (short j = 0; j < vec_size; j++) {
+        dst[i * dst_ld + j] = tmp_val[j];
+      }
+    }
+  }
+
+  /* Iteration helper */
+  METAL_FUNC void next() {
+    src += tile_stride;
+  }
+};
+
+template <int R, int C>
+struct CShape {
+  STEEL_CONST int kRows = R;
+  STEEL_CONST int kCols = C;
+};
+
+template <
+    typename T,
+    short BROWS,
+    short BCOLS,
+    short kDstStrRow,
+    short kDstStrCol,
+    short reduction_dim,
+    short tgp_size,
+    short n_reads = (BCOLS * BROWS) / (tgp_size),
+    short TCOLS = BCOLS / n_reads,
+    short TROWS = tgp_size / TCOLS>
+struct BlockLoaderT {
+  STEEL_CONST short n_rows = (BROWS + TROWS - 1) / TROWS;
+  STEEL_CONST short vec_size = n_reads;
+
+  // Leading dimension for src
+  const int src_ld;
+  const int tile_stride;
+
+  // Thread location indices
+  const short thread_idx;
+  const short bi;
+  const short bj;
+
+  // threadgroup and device memory
+  threadgroup T* dst;
+  const device T* src;
+
+  /* Constructor */
+  METAL_FUNC BlockLoaderT(
+      const device T* src_,
+      const int src_ld_,
+      threadgroup T* dst_,
+      ushort simd_group_id [[simdgroup_index_in_threadgroup]],
+      ushort simd_lane_id [[thread_index_in_simdgroup]])
+      : src_ld(src_ld_),
+        tile_stride(reduction_dim ? BCOLS : BROWS * src_ld),
+        thread_idx(simd_group_id * 32 + simd_lane_id),
+        bi(thread_idx / TCOLS),
+        bj(vec_size * (thread_idx % TCOLS)),
+        dst(dst_ + bi * kDstStrRow + bj * kDstStrCol),
+        src(src_ + bi * src_ld + bj) {}
+
+  /* Apply operation to threadgroup without bound checking */
+  template <typename UnaryOp>
+  METAL_FUNC void apply_inplace_op(thread const UnaryOp& op) const {
+    STEEL_PRAGMA_UNROLL
+    for (short i = 0; i < BROWS; i += TROWS) {
+      STEEL_PRAGMA_UNROLL
+      for (short j = 0; j < vec_size; j++) {
+        dst[i * kDstStrRow + j * kDstStrCol] =
+            op.apply(dst[i * kDstStrRow + j * kDstStrCol]);
+      }
+    }
+  }
+
+  /* Load from device memory into threadgroup memory - without bound checking */
+  METAL_FUNC void load_unsafe() const {
+    STEEL_PRAGMA_UNROLL
+    for (short i = 0; i < BROWS; i += TROWS) {
+      STEEL_PRAGMA_UNROLL
+      for (short j = 0; j < vec_size; j++) {
+        dst[i * kDstStrRow + j * kDstStrCol] = src[i * src_ld + j];
+      }
+    }
+  }
+
+  /* Load from device memory into threadgroup memory - with bound checking */
+  METAL_FUNC void load_safe(short2 src_tile_dim) const {
+    src_tile_dim = src_tile_dim - short2(bj, bi);
+
+    // Skip loading if thread has no valid reads
+    if (src_tile_dim.x <= 0 || src_tile_dim.y <= 0) {
+      STEEL_PRAGMA_UNROLL
+      for (short i = 0; i < BROWS; i += TROWS) {
+        STEEL_PRAGMA_UNROLL
+        for (short j = 0; j < vec_size; j++) {
+          dst[i * kDstStrRow + j * kDstStrCol] = T(0);
+        }
+      }
+      return;
+    }
+
+    // Use fast thread memory for bound checks
+    bool tmp_idx[vec_size];
+    T tmp_val[vec_size];
+
+    STEEL_PRAGMA_UNROLL
+    for (short i = 0; i < BROWS; i += TROWS) {
+      // Make sure tmp_idx only contains valid indices
+      STEEL_PRAGMA_UNROLL
+      for (short j = 0; j < vec_size; j++) {
+        tmp_idx[j] = (i < src_tile_dim.y) && (j < src_tile_dim.x);
+      }
+
+      // Read valid indices into tmp_val
+      STEEL_PRAGMA_UNROLL
+      for (short j = 0; j < vec_size; j++) {
+        tmp_val[j] = src[(tmp_idx[j] ? i * src_ld + j : 0)];
+      }
+
+      // Zero out uneeded values
+      STEEL_PRAGMA_UNROLL
+      for (short j = 0; j < vec_size; j++) {
+        tmp_val[j] = tmp_idx[j] ? tmp_val[j] : T(0);
+      }
+
+      // Copy values to threadgroup memory
+      STEEL_PRAGMA_UNROLL
+      for (short j = 0; j < vec_size; j++) {
+        dst[i * kDstStrRow + j * kDstStrCol] = tmp_val[j];
+      }
+    }
+  }
+
+  /* Iteration helper */
+  METAL_FUNC void next() {
+    src += tile_stride;
+  }
+};
+
+} // namespace steel
+} // namespace mlx