Implement lowering of VectorTypeCastOp to LLVM
authorNicolas Vasilache <ntv@google.com>
Fri, 18 Oct 2019 20:48:26 +0000 (13:48 -0700)
committerA. Unique TensorFlower <gardener@tensorflow.org>
Fri, 18 Oct 2019 21:00:06 +0000 (14:00 -0700)
commit2823b685804b3419c29f3fd8480f4d1ad4fb5c17
tree8e6d1707cd8b6669c70a9eaf42203c6d4fb21ec1
parent02b3ea6038b0959d5f0d62ca5ee44fbce9e3c7aa
Implement lowering of VectorTypeCastOp to LLVM

A VectorTypeCastOp can only be used to lower between statically sized contiguous memrefs of scalar and matching vector type. The sizes and strides are thus fully static and easy to determine.

A relevant test is added.

This is a step towards solving tensorflow/mlir#189.

PiperOrigin-RevId: 275538981
mlir/include/mlir/Conversion/StandardToLLVM/ConvertStandardToLLVM.h
mlir/lib/Conversion/StandardToLLVM/ConvertStandardToLLVM.cpp
mlir/lib/Conversion/VectorToLLVM/VectorToLLVM.cpp
mlir/test/Conversion/VectorToLLVM/vector-to-llvm.mlir