[llvm] r260732 - [LIR] Allow merging of memsets in negatively strided loops.

[Chad Rosier via llvm-commits]

Author: mcrosier
Date: Fri Feb 12 15:03:23 2016
New Revision: 260732

URL: http://llvm.org/viewvc/llvm-project?rev=260732&view=rev
Log:
[LIR] Allow merging of memsets in negatively strided loops.

Last part of PR25166.

Modified:
    llvm/trunk/lib/Transforms/Scalar/LoopIdiomRecognize.cpp
    llvm/trunk/test/Transforms/LoopIdiom/basic.ll

Modified: llvm/trunk/lib/Transforms/Scalar/LoopIdiomRecognize.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/LoopIdiomRecognize.cpp?rev=260732&r1=260731&r2=260732&view=diff
==============================================================================
--- llvm/trunk/lib/Transforms/Scalar/LoopIdiomRecognize.cpp (original)
+++ llvm/trunk/lib/Transforms/Scalar/LoopIdiomRecognize.cpp Fri Feb 12 15:03:23 2016
@@ -642,11 +642,12 @@ bool LoopIdiomRecognize::processLoopMemS
 
   // Check to see if the stride matches the size of the memset.  If so, then we
   // know that every byte is touched in the loop.
-  const SCEVConstant *Stride = dyn_cast<SCEVConstant>(Ev->getOperand(1));
+  const SCEVConstant *ConstStride = dyn_cast<SCEVConstant>(Ev->getOperand(1));
+  if (!ConstStride)
+    return false;
 
-  // TODO: Could also handle negative stride here someday, that will require the
-  // validity check in mayLoopAccessLocation to be updated though.
-  if (!Stride || MSI->getLength() != Stride->getValue())
+  APInt Stride = ConstStride->getAPInt();
+  if (SizeInBytes != Stride && SizeInBytes != -Stride)
     return false;
 
   // Verify that the memset value is loop invariant.  If not, we can't promote
@@ -657,9 +658,10 @@ bool LoopIdiomRecognize::processLoopMemS
 
   SmallPtrSet<Instruction *, 1> MSIs;
   MSIs.insert(MSI);
+  bool NegStride = SizeInBytes == -Stride;
   return processLoopStridedStore(Pointer, (unsigned)SizeInBytes,
                                  MSI->getAlignment(), SplatValue, MSI, MSIs, Ev,
-                                 BECount, /*NegStride=*/false);
+                                 BECount, NegStride);
 }
 
 /// mayLoopAccessLocation - Return true if the specified loop might access the

Modified: llvm/trunk/test/Transforms/LoopIdiom/basic.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/LoopIdiom/basic.ll?rev=260732&r1=260731&r2=260732&view=diff
==============================================================================
--- llvm/trunk/test/Transforms/LoopIdiom/basic.ll (original)
+++ llvm/trunk/test/Transforms/LoopIdiom/basic.ll Fri Feb 12 15:03:23 2016
@@ -531,3 +531,36 @@ for.cond.cleanup:
 ; CHECK: call void @llvm.memcpy
 ; CHECK: ret
 }
+
+; Two dimensional nested loop with negative stride should be promoted to one big memset.
+define void @test19(i8* nocapture %X) {
+entry:
+  br label %for.cond1.preheader
+
+for.cond1.preheader:                              ; preds = %entry, %for.inc4
+  %i.06 = phi i32 [ 99, %entry ], [ %dec5, %for.inc4 ]
+  %mul = mul nsw i32 %i.06, 100
+  br label %for.body3
+
+for.body3:                                        ; preds = %for.cond1.preheader, %for.body3
+  %j.05 = phi i32 [ 99, %for.cond1.preheader ], [ %dec, %for.body3 ]
+  %add = add nsw i32 %j.05, %mul
+  %idxprom = sext i32 %add to i64
+  %arrayidx = getelementptr inbounds i8, i8* %X, i64 %idxprom
+  store i8 0, i8* %arrayidx, align 1
+  %dec = add nsw i32 %j.05, -1
+  %cmp2 = icmp sgt i32 %j.05, 0
+  br i1 %cmp2, label %for.body3, label %for.inc4
+
+for.inc4:                                         ; preds = %for.body3
+  %dec5 = add nsw i32 %i.06, -1
+  %cmp = icmp sgt i32 %i.06, 0
+  br i1 %cmp, label %for.cond1.preheader, label %for.end6
+
+for.end6:                                         ; preds = %for.inc4
+  ret void
+; CHECK-LABEL: @test19(
+; CHECK: entry:
+; CHECK-NEXT: call void @llvm.memset.p0i8.i64(i8* %X, i8 0, i64 10000, i32 1, i1 false)
+; CHECK: ret void
+}