projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: db5f02b) | patch
AMDGPU/SI: Don't use reserved VGPRs for SGPR spilling
authorTom Stellard <thomas.stellard@amd.com>
Thu, 28 Jul 2016 14:30:43 +0000 (14:30 +0000)
committerTom Stellard <thomas.stellard@amd.com>
Thu, 28 Jul 2016 14:30:43 +0000 (14:30 +0000)
commit19f430109903ad56a0689de25f42295e02f9543c
treee3867e712ebc8e3a9baae3db9fefbac651fbb8b5tree | snapshot
parentdb5f02b1c48d091d0faec9db5600baacf78a843ecommit | diff
AMDGPU/SI: Don't use reserved VGPRs for SGPR spilling

Summary:
We were using reserved VGPRs for SGPR spilling and this was causing
some programs with a workgroup size of 1024 to use more than 64
registers, which is illegal.

Reviewers: arsenm, mareko, nhaehnle

Subscribers: nhaehnle, arsenm, llvm-commits, kzhuravl

Differential Revision: https://reviews.llvm.org/D22032

llvm-svn: 276980
llvm/lib/Target/AMDGPU/SIInstrInfo.cpp diff | blob | history
llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp diff | blob | history
llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp diff | blob | history
llvm/lib/Target/AMDGPU/SIRegisterInfo.h diff | blob | history
Domain: System / Toolchain;