diff options
| author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2015-03-23 18:45:30 +0000 |
|---|---|---|
| committer | Matt Arsenault <Matthew.Arsenault@amd.com> | 2015-03-23 18:45:30 +0000 |
| commit | f5b2cd891a5ed4bbddd0084cd52244b605c0b497 (patch) | |
| tree | ed373376917285f95db57fc59a0af9b955acb57e /llvm/lib/CodeGen/StackMapLivenessAnalysis.cpp | |
| parent | 05b617fed5aae1133e74e917e85e7498de0b58e9 (diff) | |
| download | bcm5719-llvm-f5b2cd891a5ed4bbddd0084cd52244b605c0b497.tar.gz bcm5719-llvm-f5b2cd891a5ed4bbddd0084cd52244b605c0b497.zip | |
R600/SI: Allow commuting compares
This enables very common cases to switch to the
smaller encoding.
All of the standard LLVM canonicalizations of comparisons
are the opposite of what we want. Compares with constants
are moved to the RHS, but the first operand can be an inline
immediate, literal constant, or SGPR using the 32-bit VOPC
encoding.
There are additional bad canonicalizations that should
also be fixed, such as canonicalizing ge x, k to gt x, (k + 1)
if this makes k no longer an inline immediate value.
llvm-svn: 232988
Diffstat (limited to 'llvm/lib/CodeGen/StackMapLivenessAnalysis.cpp')
0 files changed, 0 insertions, 0 deletions
