projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 14b947f) | patch
[clang][amdgpu] Prefer not using `fp16` conversion intrinsics.
authorMichael Liao <michael.hliao@gmail.com>
Tue, 2 Jun 2020 19:39:52 +0000 (15:39 -0400)
committerMichael Liao <michael.hliao@gmail.com>
Tue, 16 Jun 2020 14:21:56 +0000 (10:21 -0400)
commite830fa260da9d3bbe99c4176b4ddb6aa5e6229dd
tree34973d812dbd9fe03569e76ea6022831ac899889tree | snapshot
parent14b947f306ac7bc2e4eb55ac1e4255fd762b217bcommit | diff
[clang][amdgpu] Prefer not using `fp16` conversion intrinsics.

Reviewers: yaxunl, arsenm

Subscribers: kzhuravl, jvesely, wdng, nhaehnle, dstuttard, tpr, t-tye, kerbowa, cfe-commits

Tags: #clang

Differential Revision: https://reviews.llvm.org/D81849
clang/lib/Basic/Targets/AMDGPU.h diff | blob | history
clang/test/CodeGenHIP/half.hip [new file with mode: 0644] blob
Domain: System / Toolchain;