[CSSPGO]Add a flag to limit unsymbolized context depth #121531
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@llvm/pr-subscribers-pgo Author: Lei Wang (wlei-llvm) ChangesAdding a new flag( Full diff: https://github.com/llvm/llvm-project/pull/121531.diff 3 Files Affected:
diff --git a/llvm/test/tools/llvm-profgen/context-depth.test b/llvm/test/tools/llvm-profgen/context-depth.test
new file mode 100644
index 00000000000000..4eaa5fa1eae9d8
--- /dev/null
+++ b/llvm/test/tools/llvm-profgen/context-depth.test
@@ -0,0 +1,125 @@
+; Test --csprof-max-context-depth and --csprof-max-unsymbolized-context-depth
+
+; RUN: llvm-profgen --format=text --perfscript=%S/Inputs/recursion-compression-pseudoprobe.perfscript --binary=%S/Inputs/recursion-compression-pseudoprobe.perfbin --output=%t --compress-recursion=0 --profile-summary-hot-count=0 --csprof-max-context-depth=0 --csspgo-preinliner=0 --gen-cs-nested-profile=0
+; RUN: FileCheck %s --input-file %t -check-prefix=CHECK-MAX-CTX-DEPTH
+; RUN: llvm-profgen --format=text --perfscript=%S/Inputs/recursion-compression-pseudoprobe.perfscript --binary=%S/Inputs/recursion-compression-pseudoprobe.perfbin --output=%t --compress-recursion=0 --profile-summary-hot-count=0 --csprof-max-unsymbolized-context-depth=2 --csspgo-preinliner=0 --gen-cs-nested-profile=0 --skip-symbolization
+; RUN: FileCheck %s --input-file %t -check-prefix=CHECK-MAX-UNSYM-CTX-DEPTH
+; RUN: llvm-profgen --format=text --perfscript=%S/Inputs/recursion-compression-pseudoprobe.perfscript --binary=%S/Inputs/recursion-compression-pseudoprobe.perfbin --output=%t --compress-recursion=0 --profile-summary-hot-count=0 --csprof-max-unsymbolized-context-depth=2 --csspgo-preinliner=0 --gen-cs-nested-profile=0
+; RUN: FileCheck %s --input-file %t -check-prefix=CHECK-MAX-UNSYM-CTX-DEPTH-PROF
+; RUN: llvm-profgen --format=text --perfscript=%S/Inputs/recursion-compression-pseudoprobe.perfscript --binary=%S/Inputs/recursion-compression-pseudoprobe.perfbin --output=%t --compress-recursion=0 --profile-summary-hot-count=0 --csprof-max-unsymbolized-context-depth=2 --csprof-max-context-depth=0 --csspgo-preinliner=0 --gen-cs-nested-profile=0
+; RUN: FileCheck %s --input-file %t -check-prefix=CHECK-MAX-CTX-DEPTH
+
+
+; CHECK-MAX-CTX-DEPTH: [fb]:19:6
+; CHECK-MAX-CTX-DEPTH: 1: 6
+; CHECK-MAX-CTX-DEPTH: 2: 3
+; CHECK-MAX-CTX-DEPTH: 3: 3
+; CHECK-MAX-CTX-DEPTH: 4: 0
+; CHECK-MAX-CTX-DEPTH: 5: 4 fb:4
+; CHECK-MAX-CTX-DEPTH: 6: 3 fa:3
+; CHECK-MAX-CTX-DEPTH: !CFGChecksum: 563022570642068
+; CHECK-MAX-CTX-DEPTH: [fa]:14:4
+; CHECK-MAX-CTX-DEPTH: 1: 4
+; CHECK-MAX-CTX-DEPTH: 3: 4
+; CHECK-MAX-CTX-DEPTH: 4: 2
+; CHECK-MAX-CTX-DEPTH: 5: 1
+; CHECK-MAX-CTX-DEPTH: 6: 0
+; CHECK-MAX-CTX-DEPTH: 7: 2 fb:2
+; CHECK-MAX-CTX-DEPTH: 8: 1 fa:1
+; CHECK-MAX-CTX-DEPTH: !CFGChecksum: 563070469352221
+
+
+; CHECK-MAX-UNSYM-CTX-DEPTH: [0x7ab @ 0x7ab]
+; CHECK-MAX-UNSYM-CTX-DEPTH: 3
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7a0-7a7:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7a0-7ab:3
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7b2-7b5:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 3
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7a7->7b2:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7ab->7a0:4
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7b5->7c0:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: [0x7ab @ 0x7b5]
+; CHECK-MAX-UNSYM-CTX-DEPTH: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7c0-7d4:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7d4->7c0:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: [0x7b5 @ 0x7d4]
+; CHECK-MAX-UNSYM-CTX-DEPTH: 2
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7c0-7cd:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7db-7e0:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 2
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7cd->7db:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7e0->7a0:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: [0x7b5 @ 0x7e0]
+; CHECK-MAX-UNSYM-CTX-DEPTH: 2
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7a0-7a7:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7b2-7b5:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 2
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7a7->7b2:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7b5->7c0:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: [0x7d4 @ 0x7e0]
+; CHECK-MAX-UNSYM-CTX-DEPTH: 2
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7a0-7a7:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7b2-7b5:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 2
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7a7->7b2:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7b5->7c0:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: [0x7e0 @ 0x7b5]
+; CHECK-MAX-UNSYM-CTX-DEPTH: 2
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7c0-7cd:2
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7db-7e0:1
+; CHECK-MAX-UNSYM-CTX-DEPTH: 2
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7cd->7db:2
+; CHECK-MAX-UNSYM-CTX-DEPTH: 7e0->7a0:1
+
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: [fb:5 @ fb:5 @ fb]:13:4
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 1: 4
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 2: 3
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 3: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 4: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 5: 4 fb:4
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 6: 1 fa:1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: !CFGChecksum: 563022570642068
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: [fa:7 @ fb:6 @ fa]:6:2
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 1: 2
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 3: 2
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 4: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 5: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 6: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 7: 1 fb:1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 8: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: !CFGChecksum: 563070469352221
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: [fb:5 @ fb:6 @ fa]:4:1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 1: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 3: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 4: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 5: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 6: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 7: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 8: 1 fa:1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: !CFGChecksum: 563070469352221
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: [fb:6 @ fa:8 @ fa]:4:1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 1: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 3: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 4: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 5: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 6: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 7: 1 fb:1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 8: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: !CFGChecksum: 563070469352221
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: [fa:8 @ fa:7 @ fb]:3:1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 1: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 2: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 3: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 4: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 5: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 6: 1 fa:1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: !CFGChecksum: 563022570642068
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: [fb:6 @ fa:7 @ fb]:3:1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 1: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 2: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 3: 1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 4: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 5: 0
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: 6: 1 fa:1
+; CHECK-MAX-UNSYM-CTX-DEPTH-PROF: !CFGChecksum: 563022570642068
diff --git a/llvm/test/tools/llvm-profgen/recursion-compression-pseudoprobe.test b/llvm/test/tools/llvm-profgen/recursion-compression-pseudoprobe.test
index c673028584c0dc..b8e3e248e77935 100644
--- a/llvm/test/tools/llvm-profgen/recursion-compression-pseudoprobe.test
+++ b/llvm/test/tools/llvm-profgen/recursion-compression-pseudoprobe.test
@@ -9,9 +9,6 @@
; RUN: FileCheck %s --input-file %t --check-prefix=CHECK-UNWINDER
; RUN: llvm-profgen --format=text --perfscript=%S/Inputs/recursion-compression-pseudoprobe-nommap.perfscript --binary=%S/Inputs/recursion-compression-pseudoprobe.perfbin --output=%t --profile-summary-hot-count=0 --csspgo-preinliner=0 --gen-cs-nested-profile=0
; RUN: FileCheck %s --input-file %t
-; RUN: llvm-profgen --format=text --perfscript=%S/Inputs/recursion-compression-pseudoprobe.perfscript --binary=%S/Inputs/recursion-compression-pseudoprobe.perfbin --output=%t --compress-recursion=0 --profile-summary-hot-count=0 --csprof-max-context-depth=0 --csspgo-preinliner=0 --gen-cs-nested-profile=0
-; RUN: FileCheck %s --input-file %t -check-prefix=CHECK-MAX-CTX-DEPTH
-
; CHECK-UNCOMPRESS: [main:2 @ foo:5 @ fa:8 @ fa:7 @ fb:5 @ fb:5 @ fb:5 @ fb:5 @ fb:5 @ fb:5 @ fb:5 @ fb:5 @ fb:6 @ fa]:4:1
; CHECK-UNCOMPRESS: 1: 1
@@ -68,23 +65,6 @@
; CHECK-UNCOMPRESS: [main:2 @ foo:5 @ fa:8 @ fa:7 @ fb:5 @ fb:5 @ fb:5 @ fb:5 @ fb]:1:0
; CHECK-UNCOMPRESS: 5: 1 fb:1
; CHECK-UNCOMPRESS: !CFGChecksum: 563022570642068
-; CHECK-MAX-CTX-DEPTH: [fb]:19:6
-; CHECK-MAX-CTX-DEPTH: 1: 6
-; CHECK-MAX-CTX-DEPTH: 2: 3
-; CHECK-MAX-CTX-DEPTH: 3: 3
-; CHECK-MAX-CTX-DEPTH: 4: 0
-; CHECK-MAX-CTX-DEPTH: 5: 4 fb:4
-; CHECK-MAX-CTX-DEPTH: 6: 3 fa:3
-; CHECK-MAX-CTX-DEPTH: !CFGChecksum: 563022570642068
-; CHECK-MAX-CTX-DEPTH: [fa]:14:4
-; CHECK-MAX-CTX-DEPTH: 1: 4
-; CHECK-MAX-CTX-DEPTH: 3: 4
-; CHECK-MAX-CTX-DEPTH: 4: 2
-; CHECK-MAX-CTX-DEPTH: 5: 1
-; CHECK-MAX-CTX-DEPTH: 6: 0
-; CHECK-MAX-CTX-DEPTH: 7: 2 fb:2
-; CHECK-MAX-CTX-DEPTH: 8: 1 fa:1
-; CHECK-MAX-CTX-DEPTH: !CFGChecksum: 563070469352221
; CHECK: [main:2 @ foo:5 @ fa:8 @ fa:7 @ fb:5 @ fb]:13:4
diff --git a/llvm/tools/llvm-profgen/PerfReader.cpp b/llvm/tools/llvm-profgen/PerfReader.cpp
index 111c546f5329fb..9f94df6a7d9269 100644
--- a/llvm/tools/llvm-profgen/PerfReader.cpp
+++ b/llvm/tools/llvm-profgen/PerfReader.cpp
@@ -42,6 +42,11 @@ static cl::opt<bool>
cl::opt<bool> ShowDetailedWarning("show-detailed-warning",
cl::desc("Show detailed warning message."));
+static cl::opt<int> CSProfMaxUnsymbolizedCtxDepth(
+ "csprof-max-unsymbolized-context-depth", cl::init(-1),
+ cl::desc("Keep the last K contexts while merging unsymbolized profile. -1 "
+ "means no depth limit."));
+
extern cl::opt<std::string> PerfTraceFilename;
extern cl::opt<bool> ShowDisassemblyOnly;
extern cl::opt<bool> ShowSourceLocations;
@@ -172,7 +177,13 @@ std::shared_ptr<AddrBasedCtxKey> AddressStack::getContextKey() {
std::shared_ptr<AddrBasedCtxKey> KeyStr = std::make_shared<AddrBasedCtxKey>();
KeyStr->Context = Stack;
CSProfileGenerator::compressRecursionContext<uint64_t>(KeyStr->Context);
- CSProfileGenerator::trimContext<uint64_t>(KeyStr->Context);
+ int Depth = CSProfileGenerator::MaxContextDepth != -1
+ ? CSProfileGenerator::MaxContextDepth
+ : KeyStr->Context.size();
+ Depth = CSProfMaxUnsymbolizedCtxDepth != -1
+ ? std::min(static_cast<int>(CSProfMaxUnsymbolizedCtxDepth), Depth)
+ : Depth;
+ CSProfileGenerator::trimContext<uint64_t>(KeyStr->Context, Depth);
return KeyStr;
}
|
WenleiHe
reviewed
Jan 6, 2025
| ? CSProfileGenerator::MaxContextDepth | ||
| : KeyStr->Context.size(); | ||
| Depth = CSProfMaxUnsymbolizedCtxDepth != -1 | ||
| ? std::min(static_cast<int>(CSProfMaxUnsymbolizedCtxDepth), Depth) |
There was a problem hiding this comment.
Do we really need the static_cast here?
There was a problem hiding this comment.
It gave errordeduced conflicting types for parameter '_Tp' ('cl::opt<int>' vs. 'int') without the casting, it seems the deduction doesn't work very well for some environment.
Comment on lines
+180
to
+184
| int Depth = CSProfileGenerator::MaxContextDepth != -1 | ||
| ? CSProfileGenerator::MaxContextDepth | ||
| : KeyStr->Context.size(); | ||
| Depth = CSProfMaxUnsymbolizedCtxDepth != -1 | ||
| ? std::min(static_cast<int>(CSProfMaxUnsymbolizedCtxDepth), Depth) |
There was a problem hiding this comment.
It would be nice to explain the intention of having dedicated depth limit for unsymbolized context.
WenleiHe
approved these changes
Jan 7, 2025
Comment on lines
+180
to
+184
| int Depth = CSProfileGenerator::MaxContextDepth != -1 | ||
| ? CSProfileGenerator::MaxContextDepth | ||
| : KeyStr->Context.size(); | ||
| Depth = CSProfMaxUnsymbolizedCtxDepth != -1 | ||
| ? std::min(static_cast<int>(CSProfMaxUnsymbolizedCtxDepth), Depth) |
There was a problem hiding this comment.
It would be nice to explain the intention of having dedicated depth limit for unsymbolized context.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Adding a new flag(
--csprof-max-unsymbolized-context-depth) to only limit unsymbolized context depth. Currently,--csprof-max-context-depthapplies to both symbolized and unsymbolized profile context, there are scenarios where--csprof-max-context-depthmay not be flexible enough, e.g. if we want to limit the context but still preserve all the inlinings from the leaf frame(inlinings from the profiled binary), we could set the valuecsprof-max-unsymbolized-context-depth>= 1.