[llvm] [LV] Support binary and unary operations with EVL-vectorization (PR #93854)

[Florian Hahn via llvm-commits]

================
@@ -993,6 +997,64 @@ void VPWidenRecipe::execute(VPTransformState &State) {
 #endif
 }
 
+VPWidenEVLRecipe *VPWidenEVLRecipe::create(VPWidenRecipe *W, VPValue &EVL) {
+  auto *R = new VPWidenEVLRecipe(*W->getUnderlyingInstr(), W->operands(), EVL);
+  R->transferFlags(*W);
+  return R;
+}
+
+void VPWidenEVLRecipe::execute(VPTransformState &State) {
+  assert(State.UF == 1 && "Expected only UF == 1 when vectorizing with "
+                          "explicit vector length.");
+  VPValue *Op0 = getOperand(0);
+
+  // If it's scalar operation, hand translation over to VPWidenRecipe
+  if (!State.get(Op0, 0)->getType()->isVectorTy())
----------------
fhahn wrote:

should be an assert, no widen recipe should be created for scalar plans.

https://github.com/llvm/llvm-project/pull/93854