[llvm] [AArch64] Give a higher cost for more expensive SVE FCMP instructions (PR #153816)

[David Green via llvm-commits]

================
@@ -4462,6 +4490,12 @@ InstructionCost AArch64TTIImpl::getCmpSelInstrCost(
              (VecPred == FCmpInst::FCMP_ONE || VecPred == FCmpInst::FCMP_UEQ))
       Factor = 3; // fcmxx+fcmyy+or
 
+    if (isa<ScalableVectorType>(ValTy) &&
----------------
davemgreen wrote:

Yeah I have been focusing on neon and sve vla for the moment.

https://github.com/llvm/llvm-project/pull/153816