[all-commits] [llvm/llvm-project] ea8ed5: [mlir][sparse] Add F16 and BF16.

[Bixia Zheng via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: ea8ed5cbcfac973e28c9a24603395e5327a738f5
      https://github.com/llvm/llvm-project/commit/ea8ed5cbcfac973e28c9a24603395e5327a738f5
  Author: bixia1 <bixia at google.com>
  Date:   2022-06-08 (Wed, 08 Jun 2022)

  Changed paths:
    A mlir/include/mlir/ExecutionEngine/Float16bits.h
    M mlir/include/mlir/ExecutionEngine/SparseTensorUtils.h
    M mlir/lib/Dialect/SparseTensor/Transforms/CodegenUtils.cpp
    M mlir/lib/ExecutionEngine/CMakeLists.txt
    A mlir/lib/ExecutionEngine/Float16bits.cpp
    M mlir/lib/ExecutionEngine/SparseTensorUtils.cpp
    M mlir/test/Dialect/SparseTensor/conversion_sparse2dense.mlir
    A mlir/test/Integration/Dialect/SparseTensor/CPU/dense_output_f16.mlir
    A mlir/test/Integration/Dialect/SparseTensor/CPU/sparse_sum_f16.mlir
    M utils/bazel/llvm-project-overlay/mlir/BUILD.bazel

  Log Message:
  -----------
  [mlir][sparse] Add F16 and BF16.

This is the first PR to add `F16` and `BF16` support to the sparse codegen. There are still problems in supporting these two data types, such as `BF16` is not quite working yet.

Add tests cases.

Reviewed By: aartbik

Differential Revision: https://reviews.llvm.org/D127010