[PATCH] D31910: [LV] Avoid vectorizing first order recurrences when phi used outside loop

[Matthew Simpson via Phabricator via llvm-commits]

mssimpso added inline comments.


================
Comment at: lib/Transforms/Utils/LoopUtils.cpp:556-559
+  auto *Cmp = dyn_cast<CmpInst>(Latch->getTerminator()->getOperand(0));
+  bool LoopTermConditionUsesPhi = false;
+  if (Cmp && (Cmp->getOperand(0) == Phi || Cmp->getOperand(1) == Phi))
+    LoopTermConditionUsesPhi = true;
----------------
I don't think I understand this. Why do we need to check if the phi is used by the back-edge condition? If a phi is used by the back-edge condition, we should probably recognize it as an induction variable, not a first-order recurrence. I think you can remove this, and just check that the phi has no external users.


================
Comment at: test/Transforms/LoopVectorize/first-order-recurrence.ll:390-401
+for.body:
+  %inc.phi = phi i32 [ 0, %entry ], [ %inc, %for.body ]
+  %val.phi = phi i32 [ 0, %entry ], [ %addx, %for.body ]
+  %inc = add i32 %inc.phi, 1
+  %bc = zext i32 %inc.phi to i64
+  %addx = add i32 %inc.phi, %x
+  %cmp = icmp eq i32 %inc.phi, 95
----------------
In this test, "%inc.phi" is the induction variable and "%val.phi" is the potential first-order recurrence. We already allow induction variables to have external users, so we don't need to worry about this case. I think this test will pass without your patch.


https://reviews.llvm.org/D31910