lebedev.ri created this revision.
lebedev.ri added reviewers: fhahn, jdoerfert, arsenm, mkazantsev, kuhar,
brzycki, nikic.
lebedev.ri added projects: LLVM, AMDGPU.
Herald added subscribers: wenlei, kerbowa, steven_wu, hiraditya, nhaehnle,
jvesely.
lebedev.ri requested review of this revision.
Herald added subscribers: cfe-commits, wdng.
Herald added a project: clang.
My motivation is the following two problems i've hit myself:
- `FoldBranchToCommonDest()` (and possibly
`FoldValueComparisonIntoPredecessors()`) can't 'really' be extended to handle
the case where bonus instruction has external uses, in the case where we have a
successor that has multiple predecessors, but is dominated by the def block -
there may not be a PHI node, we need to form it, but to know how to fill it we
need to check for dominance, but well, we don't have the Dominator tree :)
While it would be possible to write a function to check that some block
dominates another, it seems like a hack.
(59560e85897afc50090b6c3d920bacfd28b49d06
<https://reviews.llvm.org/rG59560e85897afc50090b6c3d920bacfd28b49d06>,
https://bugs.llvm.org/show_bug.cgi?id=48450)
- Even though SimplifyCFGPass drops unreachable blocks
(`removeUnreachableBlocks()`), SimplifyCFG transformations may end up creating
more unreachable blocks, and we may happily try to SimplifyCFG them, which is,
at best, not very profitable, and at worst leads to crashes
(ef93f7a11c347534ac768ec8bbbed64cd20c41d2
<https://reviews.llvm.org/rGef93f7a11c347534ac768ec8bbbed64cd20c41d2>,
https://bugs.llvm.org/show_bug.cgi?id=48450#c8)
Indeed, simplifycfg pass didn't know how to preserve dominator tree,
it took me just under a month (starting with
e1133179587dd895962a2fe4d6eb0cb1e63b5ee2
<https://reviews.llvm.org/rGe1133179587dd895962a2fe4d6eb0cb1e63b5ee2>)
do rectify that, now it fully knows how to, there's likely some problems with
that still,
but i've dealt with everything i can spot so far.
I think we now can flip the switch.
Full disclosure: this isn't a compile-time win per-se, nor do i expect it to be
one
should we actually start to make use of DominatorTree in SimplifyCFG.
Current best-case results suggest +0.5% geomean compile-time impact:
https://llvm-compile-time-tracker.com/compare.php?from=a14c36fe27f5c36de44049237011d140a7277774&to=fd53c482aca9e20d76fb57def3ab0af851d1c697&stat=instructions
Repository:
rG LLVM Github Monorepo
https://reviews.llvm.org/D94827
Files:
clang/test/CodeGen/thinlto-distributed-newpm.ll
llvm/lib/CodeGen/DwarfEHPrepare.cpp
llvm/lib/Target/AMDGPU/AMDGPUUnifyDivergentExitNodes.cpp
llvm/lib/Transforms/Scalar/SimplifyCFGPass.cpp
llvm/lib/Transforms/Utils/SimplifyCFG.cpp
llvm/test/CodeGen/AArch64/O3-pipeline.ll
llvm/test/CodeGen/AMDGPU/opt-pipeline.ll
llvm/test/CodeGen/ARM/O3-pipeline.ll
llvm/test/Other/new-pm-defaults.ll
llvm/test/Other/new-pm-thinlto-defaults.ll
llvm/test/Other/new-pm-thinlto-postlink-pgo-defaults.ll
llvm/test/Other/new-pm-thinlto-postlink-samplepgo-defaults.ll
llvm/test/Other/new-pm-thinlto-prelink-pgo-defaults.ll
llvm/test/Other/new-pm-thinlto-prelink-samplepgo-defaults.ll
llvm/test/Other/opt-LTO-pipeline.ll
llvm/test/Other/opt-O2-pipeline.ll
llvm/test/Other/opt-O3-pipeline-enable-matrix.ll
llvm/test/Other/opt-O3-pipeline.ll
llvm/test/Other/opt-Os-pipeline.ll
llvm/test/Other/pm-pgo-preinline.ll
Index: llvm/test/Other/pm-pgo-preinline.ll
===================================================================
--- llvm/test/Other/pm-pgo-preinline.ll
+++ llvm/test/Other/pm-pgo-preinline.ll
@@ -10,7 +10,6 @@
; CHECK-Osz-NEXT: SROA
; CHECK-Osz-NEXT: Early CSE
; CHECK-Osz-NEXT: Simplify the CFG
-; CHECK-Osz-NEXT: Dominator Tree Construction
; CHECK-Osz-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-Osz-NEXT: Function Alias Analysis Results
; CHECK-Osz-NEXT: Natural Loop Information
Index: llvm/test/Other/opt-Os-pipeline.ll
===================================================================
--- llvm/test/Other/opt-Os-pipeline.ll
+++ llvm/test/Other/opt-Os-pipeline.ll
@@ -13,8 +13,8 @@
; CHECK-EXT: Good Bye World Pass
; CHECK-NOEXT-NOT: Good Bye World Pass
; CHECK-NEXT: Instrument function entry/exit with calls to e.g. mcount() (pre inlining)
-; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Dominator Tree Construction
+; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: SROA
; CHECK-NEXT: Early CSE
; CHECK-NEXT: Lower 'expect' Intrinsics
@@ -75,7 +75,6 @@
; CHECK-NEXT: Jump Threading
; CHECK-NEXT: Value Propagation
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
; CHECK-NEXT: Natural Loop Information
@@ -87,7 +86,6 @@
; CHECK-NEXT: Tail Call Elimination
; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Reassociate expressions
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Natural Loop Information
; CHECK-NEXT: Canonicalize natural loops
; CHECK-NEXT: LCSSA Verifier
@@ -104,7 +102,6 @@
; CHECK-NEXT: Loop Invariant Code Motion
; CHECK-NEXT: Unswitch loops
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
; CHECK-NEXT: Natural Loop Information
@@ -164,7 +161,6 @@
; CHECK-NEXT: Loop Pass Manager
; CHECK-NEXT: Loop Invariant Code Motion
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
; CHECK-NEXT: Natural Loop Information
@@ -232,7 +228,6 @@
; CHECK-NEXT: Optimization Remark Emitter
; CHECK-NEXT: Combine redundant instructions
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Natural Loop Information
; CHECK-NEXT: Scalar Evolution Analysis
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
Index: llvm/test/Other/opt-O3-pipeline.ll
===================================================================
--- llvm/test/Other/opt-O3-pipeline.ll
+++ llvm/test/Other/opt-O3-pipeline.ll
@@ -13,8 +13,8 @@
; CHECK-EXT: Good Bye World Pass
; CHECK-NOEXT-NOT: Good Bye World Pass
; CHECK-NEXT: Instrument function entry/exit with calls to e.g. mcount() (pre inlining)
-; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Dominator Tree Construction
+; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: SROA
; CHECK-NEXT: Early CSE
; CHECK-NEXT: Lower 'expect' Intrinsics
@@ -79,7 +79,6 @@
; CHECK-NEXT: Jump Threading
; CHECK-NEXT: Value Propagation
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Combine pattern based expressions
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
@@ -106,7 +105,6 @@
; CHECK-NEXT: Tail Call Elimination
; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Reassociate expressions
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Natural Loop Information
; CHECK-NEXT: Canonicalize natural loops
; CHECK-NEXT: LCSSA Verifier
@@ -123,7 +121,6 @@
; CHECK-NEXT: Loop Invariant Code Motion
; CHECK-NEXT: Unswitch loops
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
; CHECK-NEXT: Natural Loop Information
@@ -183,7 +180,6 @@
; CHECK-NEXT: Loop Pass Manager
; CHECK-NEXT: Loop Invariant Code Motion
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
; CHECK-NEXT: Natural Loop Information
@@ -251,7 +247,6 @@
; CHECK-NEXT: Optimization Remark Emitter
; CHECK-NEXT: Combine redundant instructions
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Natural Loop Information
; CHECK-NEXT: Scalar Evolution Analysis
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
Index: llvm/test/Other/opt-O3-pipeline-enable-matrix.ll
===================================================================
--- llvm/test/Other/opt-O3-pipeline-enable-matrix.ll
+++ llvm/test/Other/opt-O3-pipeline-enable-matrix.ll
@@ -13,8 +13,8 @@
; CHECK-EXT: Good Bye World Pass
; CHECK-NOEXT-NOT: Good Bye World Pass
; CHECK-NEXT: Instrument function entry/exit with calls to e.g. mcount() (pre inlining)
-; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Dominator Tree Construction
+; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: SROA
; CHECK-NEXT: Early CSE
; CHECK-NEXT: Lower 'expect' Intrinsics
@@ -79,7 +79,6 @@
; CHECK-NEXT: Jump Threading
; CHECK-NEXT: Value Propagation
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Combine pattern based expressions
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
@@ -106,7 +105,6 @@
; CHECK-NEXT: Tail Call Elimination
; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Reassociate expressions
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Natural Loop Information
; CHECK-NEXT: Canonicalize natural loops
; CHECK-NEXT: LCSSA Verifier
@@ -123,7 +121,6 @@
; CHECK-NEXT: Loop Invariant Code Motion
; CHECK-NEXT: Unswitch loops
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
; CHECK-NEXT: Natural Loop Information
@@ -183,7 +180,6 @@
; CHECK-NEXT: Loop Pass Manager
; CHECK-NEXT: Loop Invariant Code Motion
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
; CHECK-NEXT: Natural Loop Information
@@ -258,7 +254,6 @@
; CHECK-NEXT: Optimization Remark Emitter
; CHECK-NEXT: Combine redundant instructions
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Natural Loop Information
; CHECK-NEXT: Scalar Evolution Analysis
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
Index: llvm/test/Other/opt-O2-pipeline.ll
===================================================================
--- llvm/test/Other/opt-O2-pipeline.ll
+++ llvm/test/Other/opt-O2-pipeline.ll
@@ -13,8 +13,8 @@
; CHECK-EXT: Good Bye World Pass
; CHECK-NOEXT-NOT: Good Bye World Pass
; CHECK-NEXT: Instrument function entry/exit with calls to e.g. mcount() (pre inlining)
-; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Dominator Tree Construction
+; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: SROA
; CHECK-NEXT: Early CSE
; CHECK-NEXT: Lower 'expect' Intrinsics
@@ -75,7 +75,6 @@
; CHECK-NEXT: Jump Threading
; CHECK-NEXT: Value Propagation
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
; CHECK-NEXT: Natural Loop Information
@@ -101,7 +100,6 @@
; CHECK-NEXT: Tail Call Elimination
; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Reassociate expressions
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Natural Loop Information
; CHECK-NEXT: Canonicalize natural loops
; CHECK-NEXT: LCSSA Verifier
@@ -118,7 +116,6 @@
; CHECK-NEXT: Loop Invariant Code Motion
; CHECK-NEXT: Unswitch loops
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
; CHECK-NEXT: Natural Loop Information
@@ -178,7 +175,6 @@
; CHECK-NEXT: Loop Pass Manager
; CHECK-NEXT: Loop Invariant Code Motion
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
; CHECK-NEXT: Natural Loop Information
@@ -246,7 +242,6 @@
; CHECK-NEXT: Optimization Remark Emitter
; CHECK-NEXT: Combine redundant instructions
; CHECK-NEXT: Simplify the CFG
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Natural Loop Information
; CHECK-NEXT: Scalar Evolution Analysis
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
Index: llvm/test/Other/opt-LTO-pipeline.ll
===================================================================
--- llvm/test/Other/opt-LTO-pipeline.ll
+++ llvm/test/Other/opt-LTO-pipeline.ll
@@ -160,7 +160,6 @@
; CHECK-NEXT: Combine redundant instructions
; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Sparse Conditional Constant Propagation
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Basic Alias Analysis (stateless AA impl)
; CHECK-NEXT: Function Alias Analysis Results
; CHECK-NEXT: Natural Loop Information
@@ -182,6 +181,7 @@
; CHECK-NEXT: Lower type metadata
; CHECK-NEXT: Lower type metadata
; CHECK-NEXT: FunctionPass Manager
+; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Eliminate Available Externally Globals
; CHECK-NEXT: Dead Global Elimination
Index: llvm/test/Other/new-pm-thinlto-prelink-samplepgo-defaults.ll
===================================================================
--- llvm/test/Other/new-pm-thinlto-prelink-samplepgo-defaults.ll
+++ llvm/test/Other/new-pm-thinlto-prelink-samplepgo-defaults.ll
@@ -37,8 +37,8 @@
; CHECK-O-NEXT: Running pass: SimplifyCFGPass
; CHECK-O-NEXT: Running analysis: TargetIRAnalysis
; CHECK-O-NEXT: Running analysis: AssumptionAnalysis
-; CHECK-O-NEXT: Running pass: SROA
; CHECK-O-NEXT: Running analysis: DominatorTreeAnalysis
+; CHECK-O-NEXT: Running pass: SROA
; CHECK-O-NEXT: Running pass: EarlyCSEPass
; CHECK-O-NEXT: Running analysis: TargetLibraryAnalysis
; CHECK-O-NEXT: Running pass: LowerExpectIntrinsicPass
Index: llvm/test/Other/new-pm-thinlto-prelink-pgo-defaults.ll
===================================================================
--- llvm/test/Other/new-pm-thinlto-prelink-pgo-defaults.ll
+++ llvm/test/Other/new-pm-thinlto-prelink-pgo-defaults.ll
@@ -38,8 +38,8 @@
; CHECK-O-NEXT: Running pass: SimplifyCFGPass
; CHECK-O-NEXT: Running analysis: TargetIRAnalysis
; CHECK-O-NEXT: Running analysis: AssumptionAnalysis
-; CHECK-O-NEXT: Running pass: SROA
; CHECK-O-NEXT: Running analysis: DominatorTreeAnalysis
+; CHECK-O-NEXT: Running pass: SROA
; CHECK-O-NEXT: Running pass: EarlyCSEPass
; CHECK-O-NEXT: Running analysis: TargetLibraryAnalysis
; CHECK-O-NEXT: Running pass: LowerExpectIntrinsicPass
Index: llvm/test/Other/new-pm-thinlto-postlink-samplepgo-defaults.ll
===================================================================
--- llvm/test/Other/new-pm-thinlto-postlink-samplepgo-defaults.ll
+++ llvm/test/Other/new-pm-thinlto-postlink-samplepgo-defaults.ll
@@ -39,8 +39,8 @@
; CHECK-O-NEXT: Running pass: SimplifyCFGPass
; CHECK-O-NEXT: Running analysis: TargetIRAnalysis
; CHECK-O-NEXT: Running analysis: AssumptionAnalysis
-; CHECK-O-NEXT: Running pass: SROA
; CHECK-O-NEXT: Running analysis: DominatorTreeAnalysis
+; CHECK-O-NEXT: Running pass: SROA
; CHECK-O-NEXT: Running pass: EarlyCSEPass
; CHECK-O-NEXT: Running analysis: TargetLibraryAnalysis
; CHECK-O-NEXT: Running pass: LowerExpectIntrinsicPass
Index: llvm/test/Other/new-pm-thinlto-postlink-pgo-defaults.ll
===================================================================
--- llvm/test/Other/new-pm-thinlto-postlink-pgo-defaults.ll
+++ llvm/test/Other/new-pm-thinlto-postlink-pgo-defaults.ll
@@ -37,8 +37,8 @@
; CHECK-O-NEXT: Running pass: SimplifyCFGPass
; CHECK-O-NEXT: Running analysis: TargetIRAnalysis
; CHECK-O-NEXT: Running analysis: AssumptionAnalysis
-; CHECK-O-NEXT: Running pass: SROA
; CHECK-O-NEXT: Running analysis: DominatorTreeAnalysis
+; CHECK-O-NEXT: Running pass: SROA
; CHECK-O-NEXT: Running pass: EarlyCSEPass
; CHECK-O-NEXT: Running analysis: TargetLibraryAnalysis
; CHECK-O-NEXT: Running pass: LowerExpectIntrinsicPass
Index: llvm/test/Other/new-pm-thinlto-defaults.ll
===================================================================
--- llvm/test/Other/new-pm-thinlto-defaults.ll
+++ llvm/test/Other/new-pm-thinlto-defaults.ll
@@ -67,8 +67,8 @@
; CHECK-O-NEXT: Running pass: SimplifyCFGPass
; CHECK-O-NEXT: Running analysis: TargetIRAnalysis
; CHECK-O-NEXT: Running analysis: AssumptionAnalysis
-; CHECK-O-NEXT: Running pass: SROA
; CHECK-O-NEXT: Running analysis: DominatorTreeAnalysis
+; CHECK-O-NEXT: Running pass: SROA
; CHECK-O-NEXT: Running pass: EarlyCSEPass
; CHECK-O-NEXT: Running analysis: TargetLibraryAnalysis
; CHECK-O-NEXT: Running pass: LowerExpectIntrinsicPass
Index: llvm/test/Other/new-pm-defaults.ll
===================================================================
--- llvm/test/Other/new-pm-defaults.ll
+++ llvm/test/Other/new-pm-defaults.ll
@@ -81,8 +81,8 @@
; CHECK-O-NEXT: Running pass: SimplifyCFGPass
; CHECK-O-NEXT: Running analysis: TargetIRAnalysis
; CHECK-O-NEXT: Running analysis: AssumptionAnalysis
-; CHECK-O-NEXT: Running pass: SROA
; CHECK-O-NEXT: Running analysis: DominatorTreeAnalysis
+; CHECK-O-NEXT: Running pass: SROA
; CHECK-O-NEXT: Running pass: EarlyCSEPass
; CHECK-O-NEXT: Running analysis: TargetLibraryAnalysis
; CHECK-O-NEXT: Running pass: LowerExpectIntrinsicPass
Index: llvm/test/CodeGen/ARM/O3-pipeline.ll
===================================================================
--- llvm/test/CodeGen/ARM/O3-pipeline.ll
+++ llvm/test/CodeGen/ARM/O3-pipeline.ll
@@ -6,8 +6,8 @@
; CHECK-NEXT: Pre-ISel Intrinsic Lowering
; CHECK-NEXT: FunctionPass Manager
; CHECK-NEXT: Expand Atomic instructions
-; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Dominator Tree Construction
+; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Natural Loop Information
; CHECK-NEXT: MVE gather/scatter lowering
; CHECK-NEXT: Module Verifier
@@ -55,7 +55,6 @@
; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Exception handling preparation
; CHECK-NEXT: Merge internal globals
-; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: Natural Loop Information
; CHECK-NEXT: Scalar Evolution Analysis
; CHECK-NEXT: Lazy Branch Probability Analysis
Index: llvm/test/CodeGen/AMDGPU/opt-pipeline.ll
===================================================================
--- llvm/test/CodeGen/AMDGPU/opt-pipeline.ll
+++ llvm/test/CodeGen/AMDGPU/opt-pipeline.ll
@@ -43,8 +43,8 @@
; GCN-O1-NEXT: Function Alias Analysis Results
; GCN-O1-NEXT: Simplify well-known AMD library calls
; GCN-O1-NEXT: Instrument function entry/exit with calls to e.g. mcount() (pre inlining)
-; GCN-O1-NEXT: Simplify the CFG
; GCN-O1-NEXT: Dominator Tree Construction
+; GCN-O1-NEXT: Simplify the CFG
; GCN-O1-NEXT: SROA
; GCN-O1-NEXT: Early CSE
; GCN-O1-NEXT: Lower 'expect' Intrinsics
@@ -111,7 +111,6 @@
; GCN-O1-NEXT: Memory SSA
; GCN-O1-NEXT: Early CSE w/ MemorySSA
; GCN-O1-NEXT: Simplify the CFG
-; GCN-O1-NEXT: Dominator Tree Construction
; GCN-O1-NEXT: Basic Alias Analysis (stateless AA impl)
; GCN-O1-NEXT: Function Alias Analysis Results
; GCN-O1-NEXT: Natural Loop Information
@@ -130,7 +129,6 @@
; GCN-O1-NEXT: PGOMemOPSize
; GCN-O1-NEXT: Simplify the CFG
; GCN-O1-NEXT: Reassociate expressions
-; GCN-O1-NEXT: Dominator Tree Construction
; GCN-O1-NEXT: Natural Loop Information
; GCN-O1-NEXT: Canonicalize natural loops
; GCN-O1-NEXT: LCSSA Verifier
@@ -150,7 +148,6 @@
; GCN-O1-NEXT: Loop Pass Manager
; GCN-O1-NEXT: Unswitch loops
; GCN-O1-NEXT: Simplify the CFG
-; GCN-O1-NEXT: Dominator Tree Construction
; GCN-O1-NEXT: Basic Alias Analysis (stateless AA impl)
; GCN-O1-NEXT: Function Alias Analysis Results
; GCN-O1-NEXT: Natural Loop Information
@@ -183,7 +180,6 @@
; GCN-O1-NEXT: Post-Dominator Tree Construction
; GCN-O1-NEXT: Aggressive Dead Code Elimination
; GCN-O1-NEXT: Simplify the CFG
-; GCN-O1-NEXT: Dominator Tree Construction
; GCN-O1-NEXT: Basic Alias Analysis (stateless AA impl)
; GCN-O1-NEXT: Function Alias Analysis Results
; GCN-O1-NEXT: Natural Loop Information
@@ -250,7 +246,6 @@
; GCN-O1-NEXT: Optimization Remark Emitter
; GCN-O1-NEXT: Combine redundant instructions
; GCN-O1-NEXT: Simplify the CFG
-; GCN-O1-NEXT: Dominator Tree Construction
; GCN-O1-NEXT: Optimize scalar/vector ops
; GCN-O1-NEXT: Basic Alias Analysis (stateless AA impl)
; GCN-O1-NEXT: Function Alias Analysis Results
@@ -354,8 +349,8 @@
; GCN-O2-NEXT: Function Alias Analysis Results
; GCN-O2-NEXT: Simplify well-known AMD library calls
; GCN-O2-NEXT: Instrument function entry/exit with calls to e.g. mcount() (pre inlining)
-; GCN-O2-NEXT: Simplify the CFG
; GCN-O2-NEXT: Dominator Tree Construction
+; GCN-O2-NEXT: Simplify the CFG
; GCN-O2-NEXT: SROA
; GCN-O2-NEXT: Early CSE
; GCN-O2-NEXT: Lower 'expect' Intrinsics
@@ -428,7 +423,6 @@
; GCN-O2-NEXT: Jump Threading
; GCN-O2-NEXT: Value Propagation
; GCN-O2-NEXT: Simplify the CFG
-; GCN-O2-NEXT: Dominator Tree Construction
; GCN-O2-NEXT: Basic Alias Analysis (stateless AA impl)
; GCN-O2-NEXT: Function Alias Analysis Results
; GCN-O2-NEXT: Natural Loop Information
@@ -454,7 +448,6 @@
; GCN-O2-NEXT: Tail Call Elimination
; GCN-O2-NEXT: Simplify the CFG
; GCN-O2-NEXT: Reassociate expressions
-; GCN-O2-NEXT: Dominator Tree Construction
; GCN-O2-NEXT: Natural Loop Information
; GCN-O2-NEXT: Canonicalize natural loops
; GCN-O2-NEXT: LCSSA Verifier
@@ -474,7 +467,6 @@
; GCN-O2-NEXT: Loop Pass Manager
; GCN-O2-NEXT: Unswitch loops
; GCN-O2-NEXT: Simplify the CFG
-; GCN-O2-NEXT: Dominator Tree Construction
; GCN-O2-NEXT: Basic Alias Analysis (stateless AA impl)
; GCN-O2-NEXT: Function Alias Analysis Results
; GCN-O2-NEXT: Natural Loop Information
@@ -534,7 +526,6 @@
; GCN-O2-NEXT: Loop Pass Manager
; GCN-O2-NEXT: Loop Invariant Code Motion
; GCN-O2-NEXT: Simplify the CFG
-; GCN-O2-NEXT: Dominator Tree Construction
; GCN-O2-NEXT: Basic Alias Analysis (stateless AA impl)
; GCN-O2-NEXT: Function Alias Analysis Results
; GCN-O2-NEXT: Natural Loop Information
@@ -602,7 +593,6 @@
; GCN-O2-NEXT: Optimization Remark Emitter
; GCN-O2-NEXT: Combine redundant instructions
; GCN-O2-NEXT: Simplify the CFG
-; GCN-O2-NEXT: Dominator Tree Construction
; GCN-O2-NEXT: Natural Loop Information
; GCN-O2-NEXT: Scalar Evolution Analysis
; GCN-O2-NEXT: Basic Alias Analysis (stateless AA impl)
@@ -713,8 +703,8 @@
; GCN-O3-NEXT: Function Alias Analysis Results
; GCN-O3-NEXT: Simplify well-known AMD library calls
; GCN-O3-NEXT: Instrument function entry/exit with calls to e.g. mcount() (pre inlining)
-; GCN-O3-NEXT: Simplify the CFG
; GCN-O3-NEXT: Dominator Tree Construction
+; GCN-O3-NEXT: Simplify the CFG
; GCN-O3-NEXT: SROA
; GCN-O3-NEXT: Early CSE
; GCN-O3-NEXT: Lower 'expect' Intrinsics
@@ -791,7 +781,6 @@
; GCN-O3-NEXT: Jump Threading
; GCN-O3-NEXT: Value Propagation
; GCN-O3-NEXT: Simplify the CFG
-; GCN-O3-NEXT: Dominator Tree Construction
; GCN-O3-NEXT: Combine pattern based expressions
; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
; GCN-O3-NEXT: Function Alias Analysis Results
@@ -818,7 +807,6 @@
; GCN-O3-NEXT: Tail Call Elimination
; GCN-O3-NEXT: Simplify the CFG
; GCN-O3-NEXT: Reassociate expressions
-; GCN-O3-NEXT: Dominator Tree Construction
; GCN-O3-NEXT: Natural Loop Information
; GCN-O3-NEXT: Canonicalize natural loops
; GCN-O3-NEXT: LCSSA Verifier
@@ -838,7 +826,6 @@
; GCN-O3-NEXT: Loop Pass Manager
; GCN-O3-NEXT: Unswitch loops
; GCN-O3-NEXT: Simplify the CFG
-; GCN-O3-NEXT: Dominator Tree Construction
; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
; GCN-O3-NEXT: Function Alias Analysis Results
; GCN-O3-NEXT: Natural Loop Information
@@ -898,7 +885,6 @@
; GCN-O3-NEXT: Loop Pass Manager
; GCN-O3-NEXT: Loop Invariant Code Motion
; GCN-O3-NEXT: Simplify the CFG
-; GCN-O3-NEXT: Dominator Tree Construction
; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
; GCN-O3-NEXT: Function Alias Analysis Results
; GCN-O3-NEXT: Natural Loop Information
@@ -966,7 +952,6 @@
; GCN-O3-NEXT: Optimization Remark Emitter
; GCN-O3-NEXT: Combine redundant instructions
; GCN-O3-NEXT: Simplify the CFG
-; GCN-O3-NEXT: Dominator Tree Construction
; GCN-O3-NEXT: Natural Loop Information
; GCN-O3-NEXT: Scalar Evolution Analysis
; GCN-O3-NEXT: Basic Alias Analysis (stateless AA impl)
Index: llvm/test/CodeGen/AArch64/O3-pipeline.ll
===================================================================
--- llvm/test/CodeGen/AArch64/O3-pipeline.ll
+++ llvm/test/CodeGen/AArch64/O3-pipeline.ll
@@ -22,8 +22,8 @@
; CHECK-NEXT: FunctionPass Manager
; CHECK-NEXT: Dominator Tree Construction
; CHECK-NEXT: FunctionPass Manager
-; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Dominator Tree Construction
+; CHECK-NEXT: Simplify the CFG
; CHECK-NEXT: Natural Loop Information
; CHECK-NEXT: Lazy Branch Probability Analysis
; CHECK-NEXT: Lazy Block Frequency Analysis
Index: llvm/lib/Transforms/Utils/SimplifyCFG.cpp
===================================================================
--- llvm/lib/Transforms/Utils/SimplifyCFG.cpp
+++ llvm/lib/Transforms/Utils/SimplifyCFG.cpp
@@ -91,7 +91,7 @@
cl::opt<bool> llvm::RequireAndPreserveDomTree(
"simplifycfg-require-and-preserve-domtree", cl::Hidden, cl::ZeroOrMore,
- cl::init(false),
+ cl::init(true),
cl::desc("Temorary development switch used to gradually uplift SimplifyCFG "
"into preserving DomTree,"));
@@ -6703,20 +6703,10 @@
bool SimplifyCFGOpt::simplifyOnce(BasicBlock *BB) {
bool Changed = simplifyOnceImpl(BB);
- assert((!RequireAndPreserveDomTree ||
- (DTU &&
- DTU->getDomTree().verify(DominatorTree::VerificationLevel::Full))) &&
- "Failed to maintain validity of domtree!");
-
return Changed;
}
bool SimplifyCFGOpt::run(BasicBlock *BB) {
- assert((!RequireAndPreserveDomTree ||
- (DTU &&
- DTU->getDomTree().verify(DominatorTree::VerificationLevel::Full))) &&
- "Original domtree is invalid?");
-
bool Changed = false;
// Repeated simplify BB as long as resimplification is requested.
Index: llvm/lib/Transforms/Scalar/SimplifyCFGPass.cpp
===================================================================
--- llvm/lib/Transforms/Scalar/SimplifyCFGPass.cpp
+++ llvm/lib/Transforms/Scalar/SimplifyCFGPass.cpp
@@ -232,7 +232,7 @@
static bool simplifyFunctionCFGImpl(Function &F, const TargetTransformInfo &TTI,
DominatorTree *DT,
const SimplifyCFGOptions &Options) {
- DomTreeUpdater DTU(DT, DomTreeUpdater::UpdateStrategy::Eager);
+ DomTreeUpdater DTU(DT, DomTreeUpdater::UpdateStrategy::Lazy);
bool EverChanged = removeUnreachableBlocks(F, DT ? &DTU : nullptr);
EverChanged |= mergeEmptyReturnBlocks(F, DT ? &DTU : nullptr);
@@ -260,16 +260,8 @@
static bool simplifyFunctionCFG(Function &F, const TargetTransformInfo &TTI,
DominatorTree *DT,
const SimplifyCFGOptions &Options) {
- assert((!RequireAndPreserveDomTree ||
- (DT && DT->verify(DominatorTree::VerificationLevel::Full))) &&
- "Original domtree is invalid?");
-
bool Changed = simplifyFunctionCFGImpl(F, TTI, DT, Options);
- assert((!RequireAndPreserveDomTree ||
- (DT && DT->verify(DominatorTree::VerificationLevel::Full))) &&
- "Failed to maintain validity of domtree!");
-
return Changed;
}
Index: llvm/lib/Target/AMDGPU/AMDGPUUnifyDivergentExitNodes.cpp
===================================================================
--- llvm/lib/Target/AMDGPU/AMDGPUUnifyDivergentExitNodes.cpp
+++ llvm/lib/Target/AMDGPU/AMDGPUUnifyDivergentExitNodes.cpp
@@ -371,7 +371,7 @@
}
// FIXME: add PDT here once simplifycfg is ready.
- DomTreeUpdater DTU(DT, DomTreeUpdater::UpdateStrategy::Eager);
+ DomTreeUpdater DTU(DT, DomTreeUpdater::UpdateStrategy::Lazy);
if (RequireAndPreserveDomTree)
DTU.applyUpdates(Updates);
Updates.clear();
Index: llvm/lib/CodeGen/DwarfEHPrepare.cpp
===================================================================
--- llvm/lib/CodeGen/DwarfEHPrepare.cpp
+++ llvm/lib/CodeGen/DwarfEHPrepare.cpp
@@ -249,18 +249,8 @@
}
bool DwarfEHPrepare::run() {
- assert(((OptLevel == CodeGenOpt::None || !RequireAndPreserveDomTree) ||
- (DTU &&
- DTU->getDomTree().verify(DominatorTree::VerificationLevel::Full))) &&
- "Original domtree is invalid?");
-
bool Changed = InsertUnwindResumeCalls();
- assert(((OptLevel == CodeGenOpt::None || !RequireAndPreserveDomTree) ||
- (DTU &&
- DTU->getDomTree().verify(DominatorTree::VerificationLevel::Full))) &&
- "Original domtree is invalid?");
-
return Changed;
}
@@ -268,7 +258,7 @@
FunctionCallee &RewindFunction, Function &F,
const TargetLowering &TLI, DominatorTree *DT,
const TargetTransformInfo *TTI) {
- DomTreeUpdater DTU(DT, DomTreeUpdater::UpdateStrategy::Eager);
+ DomTreeUpdater DTU(DT, DomTreeUpdater::UpdateStrategy::Lazy);
return DwarfEHPrepare(OptLevel, RewindFunction, F, TLI, DT ? &DTU : nullptr,
TTI)
Index: clang/test/CodeGen/thinlto-distributed-newpm.ll
===================================================================
--- clang/test/CodeGen/thinlto-distributed-newpm.ll
+++ clang/test/CodeGen/thinlto-distributed-newpm.ll
@@ -34,8 +34,8 @@
; CHECK-O: Running pass: SimplifyCFGPass on main
; CHECK-O: Running analysis: TargetIRAnalysis on main
; CHECK-O: Running analysis: AssumptionAnalysis on main
-; CHECK-O: Running pass: SROA on main
; CHECK-O: Running analysis: DominatorTreeAnalysis on main
+; CHECK-O: Running pass: SROA on main
; CHECK-O: Running pass: EarlyCSEPass on main
; CHECK-O: Running analysis: TargetLibraryAnalysis on main
; CHECK-O: Running pass: LowerExpectIntrinsicPass on main
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
