[SLPVectorizer] Vectorizing the libm sqrt to llvm's sqrt intrinsic requires nnan

To quote the langref "Unlike sqrt in libm, however, llvm.sqrt has
undefined behavior for negative numbers other than -0.0 (which allows
for better optimization, because there is no need to worry about errno
being set). llvm.sqrt(-0.0) is defined to return -0.0 like IEEE sqrt."

This means that it's unsafe to replace sqrt with llvm.sqrt unless the
call is annotated with nnan.

Thanks to Hal Finkel for pointing this out!

llvm-svn: 265521

1 files changed, 3 insertions, 1 deletions

diff --git a/llvm/lib/Analysis/VectorUtils.cpp b/llvm/lib/Analysis/VectorUtils.cpp
index aa5bb7a610c..4a6cdf973e2 100644
--- a/llvm/lib/Analysis/VectorUtils.cpp
+++ b/llvm/lib/Analysis/VectorUtils.cpp
@@ -223,7 +223,9 @@ Intrinsic::ID llvm::getIntrinsicIDForCall(CallInst *CI,
   case LibFunc::sqrt:
   case LibFunc::sqrtf:
   case LibFunc::sqrtl:
-    return checkUnaryFloatSignature(*CI, Intrinsic::sqrt);
+    if (CI->hasNoNaNs())
+      return checkUnaryFloatSignature(*CI, Intrinsic::sqrt);
+    return Intrinsic::not_intrinsic;
   }
 
   return Intrinsic::not_intrinsic;