[LCSSA] Handle case with single new PHI faster.

fhahn · fhahn · commit dd2ef0af4680 · 2019-02-02T15:26:05.000Z
If there is only a single available value, all uses must be dominated by the single value and there is no need to search for a reaching definition. This drastically speeds up LCSSA in some cases. For the test case from PR37202, it speeds up LCSSA construction by 4 times. Time-passes without this patch for test case from PR37202: Total Execution Time: 29.9285 seconds (29.9276 wall clock) ---User Time--- --System Time-- --User+System-- ---Wall Time--- --- Name --- 5.2786 ( 17.7%) 0.0021 ( 1.2%) 5.2806 ( 17.6%) 5.2808 ( 17.6%) Unswitch loops 4.3739 ( 14.7%) 0.0303 ( 18.1%) 4.4042 ( 14.7%) 4.4042 ( 14.7%) Loop-Closed SSA Form Pass 4.2658 ( 14.3%) 0.0192 ( 11.5%) 4.2850 ( 14.3%) 4.2851 ( 14.3%) Loop-Closed SSA Form Pass #2 2.2307 ( 7.5%) 0.0013 ( 0.8%) 2.2320 ( 7.5%) 2.2318 ( 7.5%) Loop Invariant Code Motion 2.0888 ( 7.0%) 0.0012 ( 0.7%) 2.0900 ( 7.0%) 2.0897 ( 7.0%) Unroll loops 1.6761 ( 5.6%) 0.0013 ( 0.8%) 1.6774 ( 5.6%) 1.6774 ( 5.6%) Value Propagation 1.3686 ( 4.6%) 0.0029 ( 1.8%) 1.3716 ( 4.6%) 1.3714 ( 4.6%) Induction Variable Simplification 1.1457 ( 3.8%) 0.0010 ( 0.6%) 1.1468 ( 3.8%) 1.1468 ( 3.8%) Loop-Closed SSA Form Pass #4 1.1384 ( 3.8%) 0.0005 ( 0.3%) 1.1389 ( 3.8%) 1.1389 ( 3.8%) Loop-Closed SSA Form Pass #6 1.1360 ( 3.8%) 0.0027 ( 1.6%) 1.1387 ( 3.8%) 1.1387 ( 3.8%) Loop-Closed SSA Form Pass #5 1.1331 ( 3.8%) 0.0010 ( 0.6%) 1.1341 ( 3.8%) 1.1340 ( 3.8%) Loop-Closed SSA Form Pass #3 Time passes with this patch Total Execution Time: 19.2802 seconds (19.2813 wall clock) ---User Time--- --System Time-- --User+System-- ---Wall Time--- --- Name --- 4.4234 ( 23.2%) 0.0038 ( 2.0%) 4.4272 ( 23.0%) 4.4273 ( 23.0%) Unswitch loops 2.3828 ( 12.5%) 0.0020 ( 1.1%) 2.3848 ( 12.4%) 2.3847 ( 12.4%) Unroll loops 1.8714 ( 9.8%) 0.0020 ( 1.1%) 1.8734 ( 9.7%) 1.8735 ( 9.7%) Loop Invariant Code Motion 1.7973 ( 9.4%) 0.0022 ( 1.2%) 1.7995 ( 9.3%) 1.8003 ( 9.3%) Value Propagation 1.4010 ( 7.3%) 0.0033 ( 1.8%) 1.4043 ( 7.3%) 1.4044 ( 7.3%) Induction Variable Simplification 0.9978 ( 5.2%) 0.0244 ( 13.1%) 1.0222 ( 5.3%) 1.0224 ( 5.3%) Loop-Closed SSA Form Pass #2 0.9611 ( 5.0%) 0.0257 ( 13.8%) 0.9868 ( 5.1%) 0.9868 ( 5.1%) Loop-Closed SSA Form Pass 0.5856 ( 3.1%) 0.0015 ( 0.8%) 0.5871 ( 3.0%) 0.5869 ( 3.0%) Unroll loops #2 0.4132 ( 2.2%) 0.0012 ( 0.7%) 0.4145 ( 2.1%) 0.4143 ( 2.1%) Loop Invariant Code Motion #3 Reviewers: efriedma, davide, mzolotukhin Reviewed By: efriedma Differential Revision: https://reviews.llvm.org/D57033 llvm-svn: 352960
diff --git a/llvm/lib/Transforms/Utils/LCSSA.cpp b/llvm/lib/Transforms/Utils/LCSSA.cpp
@@ -197,6 +197,17 @@ bool llvm::formLCSSAForInstructions(SmallVectorImpl<Instruction *> &Worklist,
         continue;
       }
 
+      // If we added a single PHI, it must dominate all uses and we can directly
+      // rename it.
+      if (AddedPHIs.size() == 1) {
+        // Tell the VHs that the uses changed. This updates SCEV's caches.
+        // We might call ValueIsRAUWd multiple times for the same value.
+        if (UseToRewrite->get()->hasValueHandle())
+          ValueHandleBase::ValueIsRAUWd(*UseToRewrite, AddedPHIs[0]);
+        UseToRewrite->set(AddedPHIs[0]);
+        continue;
+      }
+
       // Otherwise, do full PHI insertion.
       SSAUpdate.RewriteUse(*UseToRewrite);
     }
@@ -210,9 +221,12 @@ bool llvm::formLCSSAForInstructions(SmallVectorImpl<Instruction *> &Worklist,
       BasicBlock *UserBB = DVI->getParent();
       if (InstBB == UserBB || L->contains(UserBB))
         continue;
-      // We currently only handle debug values residing in blocks where we have
-      // inserted a PHI instruction.
-      if (Value *V = SSAUpdate.FindValueForBlock(UserBB))
+      // We currently only handle debug values residing in blocks that were
+      // traversed while rewriting the uses. If we inserted just a single PHI,
+      // we will handle all relevant debug values.
+      Value *V = AddedPHIs.size() == 1 ? AddedPHIs[0]
+                                       : SSAUpdate.FindValueForBlock(UserBB);
+      if (V)
         DVI->setOperand(0, MetadataAsValue::get(Ctx, ValueAsMetadata::get(V)));
     }
 
diff --git a/llvm/test/Transforms/LCSSA/rewrite-existing-dbg-values.ll b/llvm/test/Transforms/LCSSA/rewrite-existing-dbg-values.ll
@@ -1,54 +1,118 @@
 ; RUN: opt -S -lcssa < %s | FileCheck %s
 
-target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
-target triple = "x86_64-unknown-linux-gnu"
-
 ; Reproducer for PR39019.
 ;
-; Verify that the llvm.dbg.value in the %for.cond.cleanup2 block is rewritten
-; to use the PHI node for %add that is created by LCSSA.
+; Verify that the llvm.dbg.values are updated to use the PHI nodes inserted by
+; LCSSA.
 
-; CHECK-LABEL: for.cond.cleanup2:
-; CHECK-NEXT: [[PN:%[^ ]*]] = phi i32 [ %add.lcssa, %for.cond.cleanup1 ]
-; CHECK-NEXT: call void @llvm.dbg.value(metadata i32 [[PN]], metadata [[VAR:![0-9]+]], metadata !DIExpression())
-; CHECK-NEXT: call void @bar(i32 [[PN]])
+; For the test case @single_exit, we can rewrite all llvm.dbg.value calls
+; to use the inserted PHI.
 
-; CHECK-LABEL: for.body:
+; CHECK-LABEL: @single_exit(
+
+; CHECK-LABEL: inner.body:
 ; CHECK: %add = add nsw i32 0, 2
-; CHECK: call void @llvm.dbg.value(metadata i32 %add, metadata [[VAR]], metadata !DIExpression())
+; CHECK: call void @llvm.dbg.value(metadata i32 %add, metadata [[VAR:![0-9]+]], metadata !DIExpression())
 
-; CHECK: [[VAR]] = !DILocalVariable(name: "sum",
 
-; Function Attrs: nounwind
-define void @foo() #0 !dbg !6 {
+; CHECK-LABEL: outer.exit:
+; CHECK-NEXT: [[PN:%[^ ]*]] = phi i32 [ %add.lcssa, %outer.latch ]
+; CHECK-NEXT: call void @llvm.dbg.value(metadata i32 [[PN]], metadata [[VAR]], metadata !DIExpression())
+; CHECK-NEXT: call void @bar(i32 [[PN]])
+
+; CHECK-LABEL: exit:
+; CHECK-NEXT: call void @llvm.dbg.value(metadata i32 [[PN]], metadata [[VAR]], metadata !DIExpression())
+
+define void @single_exit()  !dbg !6 {
 entry:
-  br label %for.cond.preheader, !dbg !12
+  br label %outer.header, !dbg !12
 
-for.cond.preheader:                               ; preds = %for.cond.cleanup1, %entry
-  br label %for.body, !dbg !12
+outer.header:                                     ; preds = %outer.latch, %entry
+  br label %inner.body, !dbg !12
 
-for.cond.cleanup2:                                ; preds = %for.cond.cleanup1
+inner.body:                                       ; preds = %inner.body, %outer.header
+  %add = add nsw i32 0, 2, !dbg !12
   call void @llvm.dbg.value(metadata i32 %add, metadata !9, metadata !DIExpression()), !dbg !12
-  tail call void @bar(i32 %add) #0, !dbg !12
-  ret void, !dbg !12
+  br i1 false, label %inner.body, label %inner.exit, !dbg !12
 
-for.cond.cleanup1:                                ; preds = %for.body
-  br i1 false, label %for.cond.preheader, label %for.cond.cleanup2, !dbg !12
+inner.exit:                                       ; preds = %inner.body
+  br label %outer.latch
 
-for.body:                                         ; preds = %for.body, %for.cond.preheader
-  %add = add nsw i32 0, 2, !dbg !12
+outer.latch:                                      ; preds = %inner.exit
+  br i1 false, label %outer.header, label %outer.exit, !dbg !12
+
+outer.exit:                                       ; preds = %outer.latch
+  call void @llvm.dbg.value(metadata i32 %add, metadata !9, metadata !DIExpression()), !dbg !12
+  tail call void @bar(i32 %add), !dbg !12
+  br label %exit
+
+exit:                                             ; preds = %outer.exit
   call void @llvm.dbg.value(metadata i32 %add, metadata !9, metadata !DIExpression()), !dbg !12
-  br i1 false, label %for.body, label %for.cond.cleanup1, !dbg !12
+  ret void, !dbg !12
 }
 
-; Function Attrs: nounwind
-declare void @bar(i32) #0
+; For the test case @multi_exit, we cannot update the llvm.dbg.value call in exit,
+; because LCSSA did not insert a PHI node in %exit, as there is no non-debug
+; use.
+
+; CHECK-LABEL: @multi_exit()
+
+; CHECK-LABEL: for.header:
+; CHECK-NEXT: %add = add nsw i32 0, 2
+; CHECK-NEXT: call void @llvm.dbg.value(metadata i32 %add, metadata [[VAR2:![0-9]+]], metadata !DIExpression())
+
+; CHECK-LABEL: for.exit1:
+; CHECK-NEXT: [[PN1:%[^ ]*]] = phi i32 [ %add, %for.header ]
+; CHECK-NEXT: br label %for.exit1.succ
+
+; CHECK-LABEL: for.exit1.succ:
+; CHECK-NEXT: call void @llvm.dbg.value(metadata i32 [[PN1]], metadata [[VAR2]], metadata !DIExpression())
+; CHECK-NEXT: call void @bar(i32 [[PN1]])
+
+; CHECK-LABEL: for.exit2:
+; CHECK-NEXT: [[PN2:%[^ ]*]] = phi i32 [ %add, %for.latch ]
+; CHECK-NEXT: call void @llvm.dbg.value(metadata i32 [[PN2]], metadata [[VAR2]], metadata !DIExpression())
+; CHECK-NEXT: call void @bar(i32 [[PN2]])
+
+; CHECK-LABEL: exit:
+; CHECK-NEXT: call void @llvm.dbg.value(metadata i32 %add, metadata [[VAR2]], metadata !DIExpression())
+
+define void @multi_exit()  !dbg !13 {
+entry:
+  br label %for.header, !dbg !14
+
+for.header:                                       ; preds = %for.latch, %entry
+  %add = add nsw i32 0, 2, !dbg !14
+  call void @llvm.dbg.value(metadata i32 %add, metadata !16, metadata !DIExpression()), !dbg !14
+  br i1 false, label %for.latch, label %for.exit1, !dbg !14
+
+for.latch:                                        ; preds = %for.header
+  br i1 false, label %for.header, label %for.exit2, !dbg !14
+
+for.exit1:                                        ; preds = %for.header
+  br label %for.exit1.succ
+
+for.exit1.succ:                                   ; preds = %for.exit1
+  call void @llvm.dbg.value(metadata i32 %add, metadata !16, metadata !DIExpression()), !dbg !14
+  tail call void @bar(i32 %add), !dbg !14
+  br label %exit
+
+for.exit2:                                        ; preds = %for.latch
+  call void @llvm.dbg.value(metadata i32 %add, metadata !16, metadata !DIExpression()), !dbg !14
+  tail call void @bar(i32 %add), !dbg !14
+  br label %exit
+
+exit:                                             ; preds = %for.exit2, %for.exit1.succ
+  call void @llvm.dbg.value(metadata i32 %add, metadata !16, metadata !DIExpression()), !dbg !14
+  ret void, !dbg !14
+}
+
+; CHECK: [[VAR]] = !DILocalVariable(name: "sum",
+; CHECK: [[VAR2]] = !DILocalVariable(name: "sum2",
 
-; Function Attrs: nounwind readnone speculatable
-declare void @llvm.dbg.value(metadata, metadata, metadata) #1
+declare void @bar(i32)
 
-attributes #0 = { nounwind }
-attributes #1 = { nounwind readnone speculatable }
+declare void @llvm.dbg.value(metadata, metadata, metadata)
 
 !llvm.dbg.cu = !{!0}
 !llvm.module.flags = !{!3, !4}
@@ -60,10 +124,14 @@ attributes #1 = { nounwind readnone speculatable }
 !3 = !{i32 2, !"Dwarf Version", i32 4}
 !4 = !{i32 2, !"Debug Info Version", i32 3}
 !5 = !{!"clang version 8.0.0"}
-!6 = distinct !DISubprogram(name: "foo", scope: !1, file: !1, line: 10, type: !7, isLocal: false, isDefinition: true, scopeLine: 10, isOptimized: true, unit: !0, retainedNodes: !8)
+!6 = distinct !DISubprogram(name: "foo", scope: !1, file: !1, line: 10, type: !7, scopeLine: 10, spFlags: DISPFlagDefinition | DISPFlagOptimized, unit: !0, retainedNodes: !8)
 !7 = !DISubroutineType(types: !2)
 !8 = !{!9}
 !9 = !DILocalVariable(name: "sum", scope: !10, file: !1, line: 11, type: !11)
 !10 = !DILexicalBlockFile(scope: !6, file: !1, discriminator: 0)
 !11 = !DIBasicType(name: "int", size: 32, encoding: DW_ATE_signed)
 !12 = !DILocation(line: 0, scope: !10)
+!13 = distinct !DISubprogram(name: "multi_exit", scope: !1, file: !1, line: 10, type: !7, scopeLine: 10, spFlags: DISPFlagDefinition | DISPFlagOptimized, unit: !0, retainedNodes: !8)
+!14 = !DILocation(line: 0, scope: !15)
+!15 = !DILexicalBlockFile(scope: !13, file: !1, discriminator: 0)
+!16 = !DILocalVariable(name: "sum2", scope: !15, file: !1, line: 11, type: !11)
