[llvm] [X86][StrictFP] Add widening support for STRICT_FMIN/STRICT_FMAX (PR #119391)

Simon Pilgrim via llvm-commits llvm-commits at lists.llvm.org
Tue Dec 10 08:12:15 PST 2024


================
@@ -33420,15 +33420,27 @@ void X86TargetLowering::ReplaceNodeResults(SDNode *N,
   case X86ISD::FMINC:
   case X86ISD::FMIN:
   case X86ISD::FMAXC:
-  case X86ISD::FMAX: {
+  case X86ISD::FMAX:
+  case X86ISD::STRICT_FMIN:
+  case X86ISD::STRICT_FMAX: {
     EVT VT = N->getValueType(0);
     assert(VT == MVT::v2f32 && "Unexpected type (!= v2f32) on FMIN/FMAX.");
+    unsigned Opc = N->getOpcode();
----------------
RKSimon wrote:

We call N->getOpcode() often enough in this function that we should probably just hoist it out at the beginning and avoid repetition/

https://github.com/llvm/llvm-project/pull/119391


More information about the llvm-commits mailing list