[PATCH] D94028: [JumpThreading][NewPM] Skip when target has divergent CF

[Arthur Eubanks via Phabricator via llvm-commits]

aeubanks created this revision.
aeubanks added a reviewer: arsenm.
Herald added a subscriber: hiraditya.
aeubanks requested review of this revision.
Herald added subscribers: llvm-commits, wdng.
Herald added a project: LLVM.

Matches the legacy pass.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D94028

Files:
  llvm/lib/Transforms/Scalar/JumpThreading.cpp
  llvm/test/Transforms/JumpThreading/divergent-target-test.ll


Index: llvm/test/Transforms/JumpThreading/divergent-target-test.ll
===================================================================
--- llvm/test/Transforms/JumpThreading/divergent-target-test.ll
+++ llvm/test/Transforms/JumpThreading/divergent-target-test.ll
@@ -1,6 +1,8 @@
 ; REQUIRES: amdgpu-registered-target && x86-registered-target
 ; RUN: opt < %s -mtriple=amdgcn -jump-threading -S | FileCheck %s  -check-prefixes=CHECK,DIVERGENT
+; RUN: opt < %s -mtriple=amdgcn -passes=jump-threading -S | FileCheck %s  -check-prefixes=CHECK,DIVERGENT
 ; RUN: opt < %s -mtriple=x86_64 -jump-threading -S | FileCheck %s  -check-prefixes=CHECK,UNIFORM
+; RUN: opt < %s -mtriple=x86_64 -passes=jump-threading -S | FileCheck %s  -check-prefixes=CHECK,UNIFORM
 
 ; Here we assure that for the target with no branch divergence usual Jump Threading optimization performed
 ; For target with branch divergence - no optimization, so the IR is unchanged.
Index: llvm/lib/Transforms/Scalar/JumpThreading.cpp
===================================================================
--- llvm/lib/Transforms/Scalar/JumpThreading.cpp
+++ llvm/lib/Transforms/Scalar/JumpThreading.cpp
@@ -341,6 +341,10 @@
 
 PreservedAnalyses JumpThreadingPass::run(Function &F,
                                          FunctionAnalysisManager &AM) {
+  auto &TTI = AM.getResult<TargetIRAnalysis>(F);
+  // Jump Threading has no sense for the targets with divergent CF
+  if (TTI.hasBranchDivergence())
+    return PreservedAnalyses::all();
   auto &TLI = AM.getResult<TargetLibraryAnalysis>(F);
   auto &DT = AM.getResult<DominatorTreeAnalysis>(F);
   auto &LVI = AM.getResult<LazyValueAnalysis>(F);


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D94028.314441.patch
Type: text/x-patch
Size: 1661 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210104/e82dcefe/attachment.bin>