[llvm] [TTI] Provide conservative costs for @llvm.speculative.load. (PR #180036)

[Florian Hahn via llvm-commits]

================
@@ -5530,6 +5530,11 @@ AArch64TTIImpl::getMemIntrinsicInstrCost(const MemIntrinsicCostAttributes &MICA,
   case Intrinsic::masked_expandload:
   case Intrinsic::masked_compressstore:
     return getMaskedMemoryOpCost(MICA, CostKind);
+  case Intrinsic::speculative_load:
+    // Scalable speculative loads are not supported yet.
+    if (isa<ScalableVectorType>(MICA.getDataType()))
+      return InstructionCost::getInvalid();
+    break;
----------------
fhahn wrote:

Removed, thanks!

https://github.com/llvm/llvm-project/pull/180036