projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 9bd5cfe) | patch
[AMDGPU] Enable 48-bit mul in AMDGPUCodeGenPrepare.
authorAbinav Puthan Purayil <abinav.puthanpurayil@amd.com>
Sat, 23 Oct 2021 15:35:32 +0000 (21:05 +0530)
committerAbinav Puthan Purayil <abinav.puthanpurayil@amd.com>
Tue, 26 Oct 2021 13:23:07 +0000 (18:53 +0530)
commit781dd39b7b6d6a3afecf7097c4fa51e194cac057
treedc1344d9a52f1572612de9c59b887c19c08ae1e5tree | snapshot
parent9bd5cfeb1fa4d7a235c1c665c7800b7891ee38a8commit | diff
[AMDGPU] Enable 48-bit mul in AMDGPUCodeGenPrepare.

We were bailing out of creating 24-bit muls for results wider than 32
bits in AMDGPUCodeGenPrepare. With the 24-bit mulhi intrinsic, this
change teaches AMDGPUCodeGenPrepare to generate the 48-bit mul
correctly.

Differential Revision: https://reviews.llvm.org/D112395
llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp diff | blob | history
llvm/test/CodeGen/AMDGPU/amdgpu-codegenprepare-mul24.ll diff | blob | history
Domain: System / Toolchain;