[PATCH] D26429: [LSR] Allow formula containing Reg for SCEVAddRecExpr with loop other than current loop

[Wei Mi via llvm-commits]

wmi updated this revision to Diff 77580.
wmi added a comment.

Updated the patch per Quentin's comment.


Repository:
  rL LLVM

https://reviews.llvm.org/D26429

Files:
  lib/Transforms/Scalar/LoopStrengthReduce.cpp
  test/Transforms/LoopStrengthReduce/nested-loop.ll


Index: lib/Transforms/Scalar/LoopStrengthReduce.cpp
===================================================================
--- lib/Transforms/Scalar/LoopStrengthReduce.cpp
+++ lib/Transforms/Scalar/LoopStrengthReduce.cpp
@@ -1048,17 +1048,16 @@
                         const Loop *L,
                         ScalarEvolution &SE, DominatorTree &DT) {
   if (const SCEVAddRecExpr *AR = dyn_cast<SCEVAddRecExpr>(Reg)) {
-    // If this is an addrec for another loop, don't second-guess its addrec phi
-    // nodes. LSR isn't currently smart enough to reason about more than one
-    // loop at a time. LSR has already run on inner loops, will not run on outer
-    // loops, and cannot be expected to change sibling loops.
+    // If this is an addrec for another loop, it should be an invariant
+    // with respect to L since L is the innermost loop (at least
+    // for now LSR only handles innermost loops).
     if (AR->getLoop() != L) {
       // If the AddRec exists, consider it's register free and leave it alone.
       if (isExistingPhi(AR, SE))
         return;
 
-      // Otherwise, do not consider this formula at all.
-      Lose();
+      // Otherwise, it will be an invariant with respect to Loop L.
+      ++NumRegs;
       return;
     }
     AddRecCost += 1; /// TODO: This should be a function of the stride.
Index: test/Transforms/LoopStrengthReduce/nested-loop.ll
===================================================================
--- test/Transforms/LoopStrengthReduce/nested-loop.ll
+++ test/Transforms/LoopStrengthReduce/nested-loop.ll
@@ -0,0 +1,63 @@
+; RUN: opt -loop-reduce -S < %s | FileCheck %s
+; Check when we use an outerloop induction variable inside of an innerloop induction value expr, LSR can still choose to use single induction variable for the innerloop and share it in multiple induction value exprs.
+
+target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-linux-gnu"
+
+define void @foo(i32 %size, i32 %nsteps, i32 %hsize, i32* %lined, i8* %maxarray) {
+entry:
+  %cmp215 = icmp sgt i32 %size, 1
+  %t0 = zext i32 %size to i64
+  %t1 = sext i32 %nsteps to i64
+  %sub2 = sub i64 %t0, 2
+  br label %for.body
+
+for.body:                                         ; preds = %for.inc, %entry
+  %indvars.iv2 = phi i64 [ %indvars.iv.next3, %for.inc ], [ 0, %entry ]
+  %t2 = mul nsw i64 %indvars.iv2, %t0
+  br i1 %cmp215, label %for.body2.preheader, label %for.inc
+
+for.body2.preheader:                              ; preds = %for.body
+  br label %for.body2
+
+; Check LSR only generates one induction variable for for.body2 and the induction variable
+; will be shared by multiple array accesses.
+; CHECK: for.body2:
+; CHECK-NEXT: [[LSR:%[^,]+]] = phi i64 [ %lsr.iv.next, %for.body2 ], [ 0, %for.body2.preheader ]
+; CHECK-NOT:  = phi i64 [ {{.*}}, %for.body2 ], [ {{.*}}, %for.body2.preheader ]
+; CHECK:      [[SCEVGEP1:%[^,]+]] = getelementptr i8, i8* %maxarray, i64 [[LSR]]
+; CHECK:      [[SCEVGEP2:%[^,]+]] = getelementptr i8, i8* [[SCEVGEP1]], i64 1
+; CHECK:      {{.*}} = load i8, i8* [[SCEVGEP2]], align 1
+; CHECK:      [[SCEVGEP3:%[^,]+]] = getelementptr i8, i8* {{.*}}, i64 [[LSR]]
+; CHECK:      {{.*}} = load i8, i8* [[SCEVGEP3]], align 1
+; CHECK:      [[SCEVGEP4:%[^,]+]] = getelementptr i8, i8* {{.*}}, i64 [[LSR]]
+; CHECK:      store i8 {{.*}}, i8* [[SCEVGEP4]], align 1
+; CHECK:      br i1 %exitcond, label %for.body2, label %for.inc.loopexit
+
+for.body2:                                        ; preds = %for.body2.preheader, %for.body2
+  %indvars.iv = phi i64 [ 1, %for.body2.preheader ], [ %indvars.iv.next, %for.body2 ]
+  %arrayidx1 = getelementptr inbounds i8, i8* %maxarray, i64 %indvars.iv
+  %v1 = load i8, i8* %arrayidx1, align 1
+  %idx2 = add nsw i64 %indvars.iv, %sub2
+  %arrayidx2 = getelementptr inbounds i8, i8* %maxarray, i64 %idx2
+  %v2 = load i8, i8* %arrayidx2, align 1
+  %tmpv = xor i8 %v1, %v2
+  %t4 = add nsw i64 %t2, %indvars.iv
+  %add.ptr = getelementptr inbounds i8, i8* %maxarray, i64 %t4
+  store i8 %tmpv, i8* %add.ptr, align 1
+  %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
+  %wide.trip.count = zext i32 %size to i64
+  %exitcond = icmp ne i64 %indvars.iv.next, %wide.trip.count
+  br i1 %exitcond, label %for.body2, label %for.inc.loopexit
+
+for.inc.loopexit:                                 ; preds = %for.body2
+  br label %for.inc
+
+for.inc:                                          ; preds = %for.inc.loopexit, %for.body
+  %indvars.iv.next3 = add nuw nsw i64 %indvars.iv2, 1
+  %cmp = icmp slt i64 %indvars.iv.next3, %t1
+  br i1 %cmp, label %for.body, label %for.end.loopexit
+
+for.end.loopexit:                                 ; preds = %for.inc
+  ret void
+}


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D26429.77580.patch
Type: text/x-patch
Size: 4725 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20161111/84d96e10/attachment.bin>