[mlir][GPUtoNVVM] Relax restriction on wmma op lowering
authorthomasraoux <thomasraoux@google.com>
Mon, 25 Oct 2021 19:42:36 +0000 (12:42 -0700)
committerthomasraoux <thomasraoux@google.com>
Thu, 28 Oct 2021 04:31:55 +0000 (21:31 -0700)
commiteacd6e1ebef511b216c8b2805e7e662b9902de74
tree32efe73f384d4247e32f8f55baef5270dc59eb9e
parentcee3419d651abdf709aff0be398fb0ac9b761723
[mlir][GPUtoNVVM] Relax restriction on wmma op lowering

Allow lowering of wmma ops with 64bits indexes. Change the default
version of the test to use default layout.

Differential Revision: https://reviews.llvm.org/D112479
mlir/lib/Conversion/GPUToNVVM/WmmaOpsToNvvm.cpp
mlir/test/Conversion/GPUToNVVM/wmma-ops-to-nvvm.mlir
mlir/test/Integration/GPU/CUDA/TensorCore/wmma-matmul-f16.mlir
mlir/test/Integration/GPU/CUDA/TensorCore/wmma-matmul-f32.mlir