Revert "[LoopVectorize] Refine runtime memory check costs when there is an outer loop"

This reverts commit a152314.

Author: david-arm

llvm/lib/Transforms/Vectorize/LoopVectorize.cpp

@@ -2055,7 +2055,7 @@ class GeneratedRTChecks {
     }
   }
 
-  InstructionCost getCost(Loop *OuterLoop) {
+  InstructionCost getCost() {
     if (SCEVCheckBlock || MemCheckBlock)
       LLVM_DEBUG(dbgs() << "Calculating cost of runtime checks:\n");
 
@@ -2076,45 +2076,16 @@ class GeneratedRTChecks {
         LLVM_DEBUG(dbgs() << "  " << C << "  for " << I << "\n");
         RTCheckCost += C;
       }
-    if (MemCheckBlock) {
-      InstructionCost MemCheckCost = 0;
+    if (MemCheckBlock)
       for (Instruction &I : *MemCheckBlock) {
         if (MemCheckBlock->getTerminator() == &I)
           continue;
         InstructionCost C =
             TTI->getInstructionCost(&I, TTI::TCK_RecipThroughput);
         LLVM_DEBUG(dbgs() << "  " << C << "  for " << I << "\n");
-        MemCheckCost += C;
-      }
-
-      // If the runtime memory checks are being created inside an outer loop
-      // we should find out if these checks are outer loop invariant. If so,
-      // the checks will be hoisted out and so the effective cost will reduce
-      // according to the outer loop trip count.
-      if (OuterLoop) {
-        ScalarEvolution *SE = MemCheckExp.getSE();
-        const SCEV *Cond = SE->getSCEV(MemRuntimeCheckCond);
-        if (SE->isLoopInvariant(Cond, OuterLoop)) {
-          if (std::optional<unsigned> OuterTC =
-                  getSmallBestKnownTC(*SE, OuterLoop))
-            MemCheckCost /= *OuterTC;
-          else {
-            // It seems reasonable to assume that we can reduce the effective
-            // cost of the checks even when we know nothing about the trip
-            // count. Here I've assumed that the outer loop executes at least
-            // twice.
-            MemCheckCost /= 2;
-          }
-
-          // Let's ensure the cost is always at least 1.
-          if (MemCheckCost == 0)
-            MemCheckCost = 1;
-        }
+        RTCheckCost += C;
       }
 
-      RTCheckCost += MemCheckCost;
-    }
-
     if (SCEVCheckBlock || MemCheckBlock)
       LLVM_DEBUG(dbgs() << "Total cost of runtime checks: " << RTCheckCost
                         << "\n");
@@ -9680,7 +9651,7 @@ static bool areRuntimeChecksProfitable(GeneratedRTChecks &Checks,
                                        std::optional<unsigned> VScale, Loop *L,
                                        ScalarEvolution &SE,
                                        ScalarEpilogueLowering SEL) {
-  InstructionCost CheckCost = Checks.getCost(L->getParentLoop());
+  InstructionCost CheckCost = Checks.getCost();
   if (!CheckCost.isValid())
     return false;
 

llvm/test/Transforms/LoopVectorize/AArch64/low_trip_memcheck_cost.ll

@@ -7,7 +7,7 @@ target triple = "aarch64-unknown-linux-gnu"
 define void @outer_no_tc(ptr nocapture noundef %a, ptr nocapture noundef readonly %b, i64 noundef %m, i64 noundef %n) {
 ; CHECK-LABEL: LV: Checking a loop in 'outer_no_tc'
 ; CHECK:      Calculating cost of runtime checks:
-; CHECK:      Total cost of runtime checks: 3
+; CHECK:      Total cost of runtime checks: 6
 ; CHECK-NEXT: LV: Minimum required TC for runtime checks to be profitable:16
 entry:
   br label %outer.loop
@@ -43,7 +43,7 @@ outer.exit:
 define void @outer_known_tc3(ptr nocapture noundef %a, ptr nocapture noundef readonly %b, i64 noundef %n) {
 ; CHECK-LABEL: LV: Checking a loop in 'outer_known_tc3'
 ; CHECK:      Calculating cost of runtime checks:
-; CHECK:      Total cost of runtime checks: 2
+; CHECK:      Total cost of runtime checks: 6
 ; CHECK-NEXT: LV: Minimum required TC for runtime checks to be profitable:16
 entry:
   br label %outer.loop
@@ -79,7 +79,7 @@ outer.exit:
 define void @outer_known_tc64(ptr nocapture noundef %a, ptr nocapture noundef readonly %b, i64 noundef %n) {
 ; CHECK-LABEL: LV: Checking a loop in 'outer_known_tc64'
 ; CHECK:      Calculating cost of runtime checks:
-; CHECK:      Total cost of runtime checks: 1
+; CHECK:      Total cost of runtime checks: 6
 ; CHECK-NEXT: LV: Minimum required TC for runtime checks to be profitable:16
 entry:
   br label %outer.loop
@@ -115,7 +115,7 @@ outer.exit:
 define void @outer_pgo_3(ptr nocapture noundef %a, ptr nocapture noundef readonly %b, i64 noundef %m, i64 noundef %n) {
 ; CHECK-LABEL: LV: Checking a loop in 'outer_pgo_3'
 ; CHECK:      Calculating cost of runtime checks:
-; CHECK:      Total cost of runtime checks: 2
+; CHECK:      Total cost of runtime checks: 6
 ; CHECK-NEXT: LV: Minimum required TC for runtime checks to be profitable:16
 entry:
   br label %outer.loop
@@ -151,8 +151,8 @@ outer.exit:
 define void @outer_known_tc3_full_range_checks(ptr nocapture noundef %dst, ptr nocapture noundef readonly %src, i64 noundef %n) {
 ; CHECK-LABEL: LV: Checking a loop in 'outer_known_tc3_full_range_checks'
 ; CHECK:      Calculating cost of runtime checks:
-; CHECK:      Total cost of runtime checks: 2
-; CHECK-NEXT: LV: Minimum required TC for runtime checks to be profitable:4
+; CHECK:      Total cost of runtime checks: 6
+; CHECK-NEXT: LV: Minimum required TC for runtime checks to be profitable:8
 entry:
   br label %outer.loop