[llvm] [VPlan] Compute induction end values in VPlan. (PR #112145)

[via llvm-commits]

================
@@ -10078,6 +10053,56 @@ LoopVectorizePass::LoopVectorizePass(LoopVectorizeOptions Opts)
       VectorizeOnlyWhenForced(Opts.VectorizeOnlyWhenForced ||
                               !EnableLoopVectorization) {}
 
+/// Prepare \p MainPlan for vectorizing the main vector loop during epilogue
+/// vectorization. Remove ResumePhis from \p MainPlan for inductions that
+/// don't have a corresponding wide induction in \p EpiPlan.
+static void preparePlanForMainVectorLoop(VPlan &MainPlan, VPlan &EpiPlan) {
+  // Collect PHI nodes of widened phis in the VPlan for the epilogue. Those
+  // will need their resume-values computed in the main vector loop. Others
+  // can be removed from the main VPlan.
+  SmallPtrSet<PHINode *, 2> EpiWidenedPhis;
+  for (VPRecipeBase &R :
+       EpiPlan.getVectorLoopRegion()->getEntryBasicBlock()->phis()) {
+    if (isa<VPCanonicalIVPHIRecipe>(&R))
+      continue;
+    EpiWidenedPhis.insert(
+        cast<PHINode>(R.getVPSingleValue()->getUnderlyingValue()));
+  }
+  for (VPRecipeBase &R : *cast<VPIRBasicBlock>(MainPlan.getScalarHeader())) {
+    auto *VPIRInst = cast<VPIRInstruction>(&R);
+    auto *IRI = dyn_cast<PHINode>(&VPIRInst->getInstruction());
+    if (!IRI)
+      break;
+    if (EpiWidenedPhis.contains(IRI))
+      continue;
+    // There is no corresponding wide induction in the epilogue plan that would
+    // need a resume value. Set the operand in VPIRInst to zero, so ResumePhi
+    // can be removed. The resume values for the scalar loop will be created
+    // during execution of EpiPlan.
+    VPRecipeBase *ResumePhi = VPIRInst->getOperand(0)->getDefiningRecipe();
+    VPIRInst->setOperand(
+        0, MainPlan.getOrAddLiveIn(Constant::getNullValue(IRI->getType())));
----------------
ayalz wrote:

> ... but in the future we can disconnect the scalar header for the main vector loop ...

When can this scalar header be disconnected? Presumably it is needed for phi's that do have wide users in EpiPlan, so could be dropped altogether (or not constructed at all) if and only if all phi's are free of such users?

>> Can VPIRInst be erased as well (perhaps with make_early_inc_range)?

> (thought I responded but cannot see the response here): VPIRinst wraps the header phi in the scalar loop, so it is probably better to keep it here.

A VPIRInst wraps an IRInst in order to (a) add operands to phi's and/or (b) support placing non-VPIRInst recipes between IRInst's, by bumping the builder. If neither is needed, VPIRInst can be dropped?
If an artificial operand is needed to appease VPIRInstruction::execute(), could undef/poison be used instead of zero?

> (There is code elsewhere that use the scalar phi to get the resume value from the main vector loop)

So dropping it or using undef would require updating said code in VPIRInstruction::execute() and/or elsewhere?

https://github.com/llvm/llvm-project/pull/112145