[llvm] [AMDGPU] Price the packed form of a vector of i1 (PR #217327)
Frederik Harwath via llvm-commits
llvm-commits at lists.llvm.org
Thu Aug 27 06:48:28 PDT 2026
================
@@ -980,13 +980,59 @@ InstructionCost GCNTTIImpl::getCFInstrCost(unsigned Opcode,
return BaseT::getCFInstrCost(Opcode, CostKind, I);
}
+// A vector of i1 has no packed form on this target: every element lives in its
+// own mask. Measured instruction counts per element to pack the masks into an
+// integer, and to unpack them again.
+static constexpr unsigned MaskPackCostPerElt = 4;
+static constexpr unsigned MaskUnpackCostPerElt = 3;
+
+/// Returns the number of elements when \p Ty is a fixed vector of i1 with more
+/// than one element.
+static std::optional<unsigned> getPackedMaskElts(Type *Ty) {
----------------
frederik-h wrote:
Rename to getNumberOfPackedMaskElts (consistent with other functions in this file) or getNumPackedMaskedElts for short.
https://github.com/llvm/llvm-project/pull/217327
More information about the llvm-commits
mailing list