================ @@ -16946,6 +16946,73 @@ SDValue SITargetLowering::performAddCombine(SDNode *N, return SDValue(); } +// Try to fold saturating add with dot product intrinsic into dot instruction +// with clamp. Matches patterns like: +// uaddsat(dot4(..., 0), c) -> dot4(..., c) clamp +// uaddsat(dot2(..., 0), c) -> dot2(..., c) clamp +SDValue SITargetLowering::performSatAddCombine(SDNode *N, ---------------- arsenm wrote:
Is this version still tested? https://github.com/llvm/llvm-project/pull/187945 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
