[llvm] [LV] Unify interleaved load handling for fixed and scalable VFs. nfc (PR #146914)

[Florian Hahn via llvm-commits]

================
@@ -3490,57 +3490,39 @@ void VPInterleaveRecipe::execute(VPTransformState &State) {
     Group->addMetadata(NewLoad);
 
     ArrayRef<VPValue *> VPDefs = definedValues();
-    const DataLayout &DL = State.CFG.PrevBB->getDataLayout();
     if (VecTy->isScalableTy()) {
       // Scalable vectors cannot use arbitrary shufflevectors (only splats),
       // so must use intrinsics to deinterleave.
       assert(InterleaveFactor <= 8 &&
              "Unsupported deinterleave factor for scalable vectors");
-      Value *Deinterleave = State.Builder.CreateIntrinsic(
+      NewLoad = State.Builder.CreateIntrinsic(
           getDeinterleaveIntrinsicID(InterleaveFactor), NewLoad->getType(),
           NewLoad,
           /*FMFSource=*/nullptr, "strided.vec");
+    }
 
-      for (unsigned I = 0, J = 0; I < InterleaveFactor; ++I) {
-        Instruction *Member = Group->getMember(I);
-        Value *StridedVec = State.Builder.CreateExtractValue(Deinterleave, I);
-        if (!Member) {
-          // This value is not needed as it's not used
-          cast<Instruction>(StridedVec)->eraseFromParent();
----------------
fhahn wrote:

This should be fine, in the new version we are skipping members before creating the strided vector, which could be done here as well instead of erasing if there's no member I think

https://github.com/llvm/llvm-project/pull/146914