diff options
| author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2020-02-10 10:30:34 -0500 |
|---|---|---|
| committer | Matt Arsenault <arsenm2@gmail.com> | 2020-02-11 18:35:00 -0500 |
| commit | 86f9117d476bcef2f5e0eabae4781e99877ce7b5 (patch) | |
| tree | 0928e0493bbb31d255ed9bce634c8f15626f3aae /llvm/lib/Support/ErrorHandling.cpp | |
| parent | b2c44de956cca22efa374cfb587912b38c41ed67 (diff) | |
| download | llvm-86f9117d476bcef2f5e0eabae4781e99877ce7b5.zip llvm-86f9117d476bcef2f5e0eabae4781e99877ce7b5.tar.gz llvm-86f9117d476bcef2f5e0eabae4781e99877ce7b5.tar.bz2 | |
AMDGPU: Don't report 2-byte alignment as fast
This is apparently worse than 1-byte alignment. This does not attempt
to decompose 2-byte aligned wide stores, but will stop trying to
produce them.
Also fix bug in LoadStoreVectorizer which was decreasing the alignment
and vectorizing stack accesses. It was assuming a stack object was an
alloca that could have its base alignment changed, which is not true
if the pointer is derived from a function argument.
Diffstat (limited to 'llvm/lib/Support/ErrorHandling.cpp')
0 files changed, 0 insertions, 0 deletions