[llvm] [SLP] Enable optimization of freeze instructions (PR #102217)

Alexey Bataev via llvm-commits llvm-commits at lists.llvm.org
Tue Aug 6 14:25:09 PDT 2024


================
@@ -9790,10 +9792,11 @@ BoUpSLP::getEntryCost(const TreeEntry *E, ArrayRef<Value *> VectorizedVals,
   case Instruction::AShr:
   case Instruction::And:
   case Instruction::Or:
-  case Instruction::Xor: {
+  case Instruction::Xor:
+  case Instruction::Freeze: {
----------------
alexey-bataev wrote:

I rather doubt it should be here. I think the cost of freeze must be 0 always

https://github.com/llvm/llvm-project/pull/102217


More information about the llvm-commits mailing list