JIT compile option for binary minimization (#1091)

* try cpp 20 for compile

* unary, binary, ternary in jit

* nits

* fix gather/scatter

* fix rebase

* reorg compile

* add ternary to compile

* jit copy

* jit compile flag

* fix build

* use linked function for ternary

* some nits

* docs + circle min size build

* docs + circle min size build

* fix extension

* fix no cpu build

* improve includes

mlx/backend/metal/nojit_kernels.cpp

@@ -0,0 +1,46 @@
+
+#include "mlx/backend/metal/kernels.h"
+#include "mlx/backend/metal/utils.h"
+#include "mlx/primitives.h"
+
+namespace mlx::core {
+
+MTL::ComputePipelineState* get_unary_kernel(
+    metal::Device& d,
+    const std::string& kernel_name,
+    const array&) {
+  return d.get_kernel(kernel_name);
+}
+
+MTL::ComputePipelineState* get_binary_kernel(
+    metal::Device& d,
+    const std::string& kernel_name,
+    const array& in,
+    const array& out) {
+  return d.get_kernel(kernel_name);
+}
+
+MTL::ComputePipelineState* get_binary_two_kernel(
+    metal::Device& d,
+    const std::string& kernel_name,
+    const array& in,
+    const array& out) {
+  return d.get_kernel(kernel_name);
+}
+
+MTL::ComputePipelineState* get_ternary_kernel(
+    metal::Device& d,
+    const std::string& kernel_name,
+    const array& out) {
+  return d.get_kernel(kernel_name);
+}
+
+MTL::ComputePipelineState* get_copy_kernel(
+    metal::Device& d,
+    const std::string& kernel_name,
+    const array& in,
+    const array& out) {
+  return d.get_kernel(kernel_name);
+}
+
+} // namespace mlx::core