[LAA-memchecks 2/3] Move number of memcheck threshold checking to LV

Now the analysis won't "fail" if the memchecks exceed the threshold.  It
is the transform pass' responsibility to perform the check.

This allows the transform pass to further analyze/eliminate the
memchecks.  E.g. in Loop distribution we only need to check pointers
that end up in different partitions.

Note that there is a slight change of functionality here.  The logic in
analyzeLoop is that if dependence checking fails due to non-constant
distance between the pointers, another attempt is made to prove safety
of the dependences purely using run-time checks.

Before this patch we could fail the loop due to exceeding the memcheck
threshold after the first step, now we only check the threshold in the
client after the full analysis.  There is no measurable compile-time
effect but I wanted to record this here.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231817 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/include/llvm/Analysis/LoopAccessAnalysis.h b/include/llvm/Analysis/LoopAccessAnalysis.h
index 31b0b8eb21dd31673d9b465c80827535853c0387..d188327fbf4fcdfca89c4e8cf85096b342c44c36 100644 (file)
--- a/include/llvm/Analysis/LoopAccessAnalysis.h
+++ b/include/llvm/Analysis/LoopAccessAnalysis.h
@@ -362,6 +362,10 @@ public:
     return &PtrRtCheck;
   }
 
+  /// \brief Number of memchecks required to prove independence of otherwise
+  /// may-alias pointers.
+  unsigned getNumRuntimePointerChecks() const { return NumComparisons; }
+
   /// Return true if the block BB needs to be predicated in order for the loop
   /// to be vectorized.
   static bool blockNeedsPredication(BasicBlock *BB, Loop *TheLoop,
@@ -416,6 +420,10 @@ private:
   /// loop-independent and loop-carried dependences between memory accesses.
   MemoryDepChecker DepChecker;
 
+  /// \brief Number of memchecks required to prove independence of otherwise
+  /// may-alias pointers
+  unsigned NumComparisons;
+
   Loop *TheLoop;
   ScalarEvolution *SE;
   const DataLayout &DL;

diff --git a/lib/Analysis/LoopAccessAnalysis.cpp b/lib/Analysis/LoopAccessAnalysis.cpp
index 27ccd66541fce4b00b47ad16f144a871da941ddc..d044bb008e99cf82d638f0e5d3301370c29d8005 100644 (file)
--- a/lib/Analysis/LoopAccessAnalysis.cpp
+++ b/lib/Analysis/LoopAccessAnalysis.cpp
@@ -1072,7 +1072,6 @@ void LoopAccessInfo::analyzeLoop(const ValueToValueMap &Strides) {
 
   // Find pointers with computable bounds. We are going to use this information
   // to place a runtime bound check.
-  unsigned NumComparisons = 0;
   bool CanDoRT = false;
   if (NeedRTCheck)
     CanDoRT = Accesses.canCheckPtrAtRT(PtrRtCheck, NumComparisons, SE, TheLoop,
@@ -1098,17 +1097,6 @@ void LoopAccessInfo::analyzeLoop(const ValueToValueMap &Strides) {
     return;
   }
 
-  if (NumComparisons > RuntimeMemoryCheckThreshold) {
-    emitAnalysis(LoopAccessReport()
-                 << NumComparisons << " exceeds limit of "
-                 << RuntimeMemoryCheckThreshold
-                 << " dependent memory operations checked at runtime");
-    DEBUG(dbgs() << "LAA: Too many memory checks needed.\n");
-    PtrRtCheck.reset();
-    CanVecMem = false;
-    return;
-  }
-
   PtrRtCheck.Need = NeedRTCheck;
 
   CanVecMem = true;
@@ -1140,18 +1128,6 @@ void LoopAccessInfo::analyzeLoop(const ValueToValueMap &Strides) {
         return;
       }
 
-      // Check that we did not collect too many pointers.
-      if (NumComparisons > RuntimeMemoryCheckThreshold) {
-        emitAnalysis(LoopAccessReport()
-                     << NumComparisons << " exceeds limit of "
-                     << RuntimeMemoryCheckThreshold
-                     << " dependent memory operations checked at runtime");
-        DEBUG(dbgs() << "LAA: Can't vectorize with memory checks\n");
-        PtrRtCheck.reset();
-        CanVecMem = false;
-        return;
-      }
-
       CanVecMem = true;
     }
   }
@@ -1283,9 +1259,9 @@ LoopAccessInfo::LoopAccessInfo(Loop *L, ScalarEvolution *SE,
                                const TargetLibraryInfo *TLI, AliasAnalysis *AA,
                                DominatorTree *DT,
                                const ValueToValueMap &Strides)
-    : DepChecker(SE, L), TheLoop(L), SE(SE), DL(DL), TLI(TLI), AA(AA),
-      DT(DT), NumLoads(0), NumStores(0), MaxSafeDepDistBytes(-1U),
-      CanVecMem(false) {
+    : DepChecker(SE, L), NumComparisons(0), TheLoop(L), SE(SE), DL(DL),
+      TLI(TLI), AA(AA), DT(DT), NumLoads(0), NumStores(0),
+      MaxSafeDepDistBytes(-1U), CanVecMem(false) {
   if (canAnalyzeLoop())
     analyzeLoop(Strides);
 }

diff --git a/lib/Transforms/Vectorize/LoopVectorize.cpp b/lib/Transforms/Vectorize/LoopVectorize.cpp
index 18a456f611aea29ebe74b7a3f56862c1d6f18554..1abefd75c97ab2fcaafb699a5f492828a096830e 100644 (file)
--- a/lib/Transforms/Vectorize/LoopVectorize.cpp
+++ b/lib/Transforms/Vectorize/LoopVectorize.cpp
@@ -3874,7 +3874,19 @@ bool LoopVectorizationLegality::canVectorizeMemory() {
   auto &OptionalReport = LAI->getReport();
   if (OptionalReport)
     emitAnalysis(VectorizationReport(*OptionalReport));
-  return LAI->canVectorizeMemory();
+  if (!LAI->canVectorizeMemory())
+    return false;
+
+  if (LAI->getNumRuntimePointerChecks() >
+      VectorizerParams::RuntimeMemoryCheckThreshold) {
+    emitAnalysis(VectorizationReport()
+                 << LAI->getNumRuntimePointerChecks() << " exceeds limit of "
+                 << VectorizerParams::RuntimeMemoryCheckThreshold
+                 << " dependent memory operations checked at runtime");
+    DEBUG(dbgs() << "LV: Too many memory checks needed.\n");
+    return false;
+  }
+  return true;
 }
 
 static bool hasMultipleUsesOf(Instruction *I,