Start to cleanup/unify accelerate and common back-ends (Part 1/N) (#1777)

* start to cleanup/unify accelerate and common back-ends

* more progress

* simplify

* add half type and allow infs in simd exp

* unify softmax + quantized, more dispatches to simd quantized mm

* add sin/cos, use simd in vector-scalar ops

* faster CPU vectorize quant

* faster erf/erfinv

mlx/backend/common/load.cpp

@@ -1,12 +1,9 @@
 // Copyright © 2023 Apple Inc.
 
 #include <algorithm>
-#include <cassert>
 #include <utility>
 
-#include "mlx/allocator.h"
 #include "mlx/backend/common/load.h"
-#include "mlx/primitives.h"
 
 namespace {
 
@@ -51,11 +48,4 @@ void load(
   }
 }
 
-void Load::eval(const std::vector<array>& inputs, array& out) {
-  assert(inputs.size() == 0);
-  out.set_data(allocator::malloc_or_wait(out.nbytes()));
-
-  load(out, offset_, reader_, swap_endianness_);
-}
-
 } // namespace mlx::core