[llvm] [LV] Ignore some costs when loop gets fully unrolled (PR #106699)

[David Sherwood via llvm-commits]

================
@@ -7223,6 +7251,17 @@ LoopVectorizationPlanner::precomputeCosts(VPlan &Plan, ElementCount VF,
         continue;
       IVInsts.push_back(CI);
     }
+
+    // If the vector loop gets executed exactly once with the given VF, ignore
+    // the costs of comparison and induction instructions, as they'll get
+    // simplified away.
+    // TODO: Remove this code after stepping away from the legacy cost model and
+    // adding code to simplify VPlans before calculating their costs.
+    auto TC = PSE.getSE()->getSmallConstantTripCount(OrigLoop);
+    if (VF.isFixed() && TC == VF.getFixedValue())
----------------
david-arm wrote:

I think this should check we're not tail-folding here, i.e.

`if (VF.isFixed() && TC == VF.getFixedValue() && !CM.foldTailByMasking())`

https://github.com/llvm/llvm-project/pull/106699