projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 801c2b9) | patch
[AMDGPU] Update gfx90a memory model support
authorTony Tye <Tony.Tye@amd.com>
Fri, 7 May 2021 20:55:23 +0000 (20:55 +0000)
committerTony Tye <Tony.Tye@amd.com>
Wed, 30 Jun 2021 04:05:22 +0000 (04:05 +0000)
commit7f19aa73c265b966a5ecac4abe78b862426e28b5
treea484fb13461fa6ad3d6da5457b0b19609621ab4dtree | snapshot
parent801c2b9bbaad778fd4f9fb25b4ab2bd8742a5a3bcommit | diff
[AMDGPU] Update gfx90a memory model support

Update AMDGPU gfx90a memory model to make coarse grain memory allocations
consistent when fine grained system scope atomic acquire and release is
performed.

Reviewed By: rampitec

Differential Revision: https://reviews.llvm.org/D105137
llvm/docs/AMDGPUUsage.rst diff | blob | history
llvm/lib/Target/AMDGPU/SIMemoryLegalizer.cpp diff | blob | history
llvm/test/CodeGen/AMDGPU/fp64-atomics-gfx90a.ll diff | blob | history
llvm/test/CodeGen/AMDGPU/global-atomics-fp.ll diff | blob | history
llvm/test/CodeGen/AMDGPU/memory-legalizer-fence.ll diff | blob | history
llvm/test/CodeGen/AMDGPU/memory-legalizer-flat-system.ll diff | blob | history
llvm/test/CodeGen/AMDGPU/memory-legalizer-global-system.ll diff | blob | history
Domain: System / Toolchain;