RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-23 11:13:28 +01:00
Code Issues Projects Releases Wiki Activity

LoopVersioning: Respect convergent

Browse Source 
This changes the standalone pass only. Arguably the utility class
itself should assert there are no convergent calls. However, a target
pass with additional context may still be able to version a loop if
all of the dynamic conditions are sufficiently uniform.

llvm-svn: 363165
This commit is contained in:
Matt Arsenault 2019-06-12 14:05:58 +00:00
parent 5b531e85df
commit 1f39fe8ef2
2 changed files with 43 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
lib/Transforms/Utils/LoopVersioning.cpp
View File

@ -280,7 +280,8 @@ public:
bool Changed = false;
for (Loop *L : Worklist) {
const LoopAccessInfo &LAI = LAA->getInfo(L);
if (L->isLoopSimplifyForm() && (LAI.getNumRuntimePointerChecks() ||
if (L->isLoopSimplifyForm() && !LAI.hasConvergentOp() &&
(LAI.getNumRuntimePointerChecks() ||
!LAI.getPSE().getUnionPredicate().isAlwaysTrue())) {
LoopVersioning LVer(LAI, L, LI, DT, SE);
LVer.versionLoop();

40
test/Transforms/LoopVersioning/convergent.ll Normal file
View File

@ -0,0 +1,40 @@
; RUN: opt -basicaa -loop-versioning -S < %s | FileCheck %s
target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128"
; Do not version this loop because of a convergent operation
; CHECK-LABEL: @f(
; CHECK: call i32 @llvm.convergent(
; CHECK-NOT: call i32 @llvm.convergent(
define void @f(i32* %a, i32* %b, i32* %c) #0 {
entry:
br label %for.body
for.body: ; preds = %for.body, %entry
%ind = phi i64 [ 0, %entry ], [ %add, %for.body ]
%arrayidxA = getelementptr inbounds i32, i32* %a, i64 %ind
%loadA = load i32, i32* %arrayidxA, align 4
%arrayidxB = getelementptr inbounds i32, i32* %b, i64 %ind
%loadB = load i32, i32* %arrayidxB, align 4
%convergentB = call i32 @llvm.convergent(i32 %loadB)
%mulC = mul i32 %loadA, %convergentB
%arrayidxC = getelementptr inbounds i32, i32* %c, i64 %ind
store i32 %mulC, i32* %arrayidxC, align 4
%add = add nuw nsw i64 %ind, 1
%exitcond = icmp eq i64 %add, 20
br i1 %exitcond, label %for.end, label %for.body
for.end: ; preds = %for.body
ret void
}
declare i32 @llvm.convergent(i32) #1
attributes #0 = { nounwind convergent }
attributes #1 = { nounwind readnone convergent }