2 Commits

Author SHA1 Message Date
Matt Arsenault 45cd9c275c AMDGPU: Use SGPR_128 instead of SReg_128 for vregs
SGPR_128 only includes the real allocatable SGPRs, and SReg_128 adds
the additional non-allocatable TTMP registers. There's no point in
allocating SReg_128 vregs. This shrinks the size of the classes
regalloc needs to consider, which is usually good.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@374284 91177308-0d34-0410-b5e6-96231b3b80d8
2019-10-10 07:11:33 +00:00
Stanislav Mekhanoshin 4b0f16838a [AMDGPU] gfx1010 subvector test. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@363623 91177308-0d34-0410-b5e6-96231b3b80d8
2019-06-17 21:55:06 +00:00