[llvm] [NVPTX] Improve modeling of inline PTX (PR #130675)

[Alex MacLean via llvm-commits]

================
@@ -483,6 +484,34 @@ NVPTXTTIImpl::instCombineIntrinsic(InstCombiner &IC, IntrinsicInst &II) const {
   return std::nullopt;
 }
 
+InstructionCost
+NVPTXTTIImpl::getInstructionCost(const User *U,
+                                 ArrayRef<const Value *> Operands,
+                                 TTI::TargetCostKind CostKind) {
+  if (const auto *CI = dyn_cast<CallInst>(U))
+    if (const auto *IA = dyn_cast<InlineAsm>(CI->getCalledOperand())) {
+      // Without this implementation getCallCost() would return the number
+      // of arguments+1 as the cost. Because the cost-model assumes it is a call
+      // since it is classified as a call in the IR. A better cost model would
+      // be to return the number of asm instructions embedded in the asm
+      // string.
+      auto &AsmStr = IA->getAsmString();
+      SmallVector<StringRef, 4> AsmPieces;
+      SplitString(AsmStr, AsmPieces, ";\n");
----------------
AlexMaclean wrote:

Yea, that is true. No matter what, there are going to be edge cases that are not correctly estimated here. I've switched to splitting only on ";" but my assumption is this hasn't fixed all the examples you listed above. I do think it will be good enough for most inline PTX though.

https://github.com/llvm/llvm-project/pull/130675