RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-18 18:42:46 +02:00
Code Issues Projects Releases Wiki Activity

AMDGPU: Put inexpensive ops first in AMDGPUAnnotateUniformValues::visitLoadInst

Browse Source 
Summary:
  This is in response to the review of https://reviews.llvm.org/D84873:
The expensive check should be reordered last

Reviewers:
  arsenm

Differential Revision:
  https://reviews.llvm.org/D84890

(cherry picked from commit 243376cdc7b719d443f42c8c4667e5d96af53dcc)
This commit is contained in:
Changpeng Fang 2020-07-30 14:37:06 -07:00 committed by Hans Wennborg
parent 2fc661ffb0
commit 5aeae1780f
1 changed files with 13 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
lib/Target/AMDGPU/AMDGPUAnnotateUniformValues.cpp
View File

@ -131,10 +131,20 @@ void AMDGPUAnnotateUniformValues::visitLoadInst(LoadInst &I) {
// We're tracking up to the Function boundaries, and cannot go beyond because
// of FunctionPass restrictions. We can ensure that is memory not clobbered
// for memory operations that are live in to entry points only.
bool NotClobbered = isEntryFunc && !isClobberedInFunction(&I);
Instruction *PtrI = dyn_cast<Instruction>(Ptr);
if (!PtrI && NotClobbered && isGlobalLoad(I)) {
if (isa<Argument>(Ptr) || isa<GlobalValue>(Ptr)) {
if (!isEntryFunc) {
if (PtrI)
setUniformMetadata(PtrI);
return;
}
bool NotClobbered = false;
if (PtrI)
NotClobbered = !isClobberedInFunction(&I);
else if (isa<Argument>(Ptr) || isa<GlobalValue>(Ptr)) {
if (isGlobalLoad(I) && !isClobberedInFunction(&I)) {
NotClobbered = true;
// Lookup for the existing GEP
if (noClobberClones.count(Ptr)) {
PtrI = noClobberClones[Ptr];