[llvm] [VPlan] First step towards VPlan cost modeling. (PR #67934)

Florian Hahn via llvm-commits llvm-commits at lists.llvm.org
Fri May 10 07:43:36 PDT 2024


================
@@ -1371,8 +1387,6 @@ class VPWidenCastRecipe : public VPRecipeWithIRFlags {
         ResultTy(ResultTy) {
     assert(UI.getOpcode() == Opcode &&
            "opcode of underlying cast doesn't match");
-    assert(UI.getType() == ResultTy &&
-           "result type of underlying cast doesn't match");
----------------
fhahn wrote:

No, as we retain the underlying instruction in a narrower version of the cast, so we can still query the cost model for the underlying instruction, even after VP2VP narrowing. This is needed until we handle cast-costs completely in VPlan.

https://github.com/llvm/llvm-project/pull/67934


More information about the llvm-commits mailing list