[LoopVectorize] Preserve SCEV

As far as I can tell, LoopVectorize preserves SCEV, mainly by dint
of forgetting the loop being vectorized. We should mark it as
preserved in the pass manager.

This is a very small compile-time improvement.

Differential Revision: https://reviews.llvm.org/D149147

diff --git a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
index 645b62d..c9fd816 100644 (file)
--- a/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
+++ b/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp
@@ -10610,6 +10610,11 @@ PreservedAnalyses LoopVectorizePass::run(Function &F,
     if (!EnableVPlanNativePath) {
       PA.preserve<LoopAnalysis>();
       PA.preserve<DominatorTreeAnalysis>();
+      PA.preserve<ScalarEvolutionAnalysis>();
+
+#ifdef EXPENSIVE_CHECKS
+      SE.verify();
+#endif
     }
 
     if (Result.MadeCFGChange) {

diff --git a/llvm/test/Transforms/LoopVectorize/novect-lcssa-cfg-invalidation.ll b/llvm/test/Transforms/LoopVectorize/novect-lcssa-cfg-invalidation.ll
index 24bcfe0..c78e005 100644 (file)
--- a/llvm/test/Transforms/LoopVectorize/novect-lcssa-cfg-invalidation.ll
+++ b/llvm/test/Transforms/LoopVectorize/novect-lcssa-cfg-invalidation.ll
@@ -8,7 +8,7 @@ target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
 define i32 @novect(ptr %p) {
 
 ; CHECK:           Running pass: LoopVectorizePass on novect
-; CHECK:           Invalidating analysis: ScalarEvolutionAnalysis on novect
+; CHECK-NOT:       Invalidating analysis: ScalarEvolutionAnalysis on novect
 ; CHECK-NOT:       Invalidating analysis: BranchProbabilityAnalysis on novect
 ; CHECK-NOT:       Invalidating analysis: BlockFrequencyAnalysis on novect
 ; CHECK:           Invalidating analysis: DemandedBitsAnalysis on novect