[memprof] Add RecordSerializationRoundTripVersion4HotColdSchema #156783
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
I'm planning to remove the V2 support. Now, HotColdSchema is one feature that's tested in V2 but not in V4. This patch derives: RecordSerializationRoundTripVersion4HotColdSchema from: RecordSerializationRoundTripVersion2HotColdSchema and then adds DummyMap, a V4-speicifc item.
|
@llvm/pr-subscribers-pgo Author: Kazu Hirata (kazutakahirata) ChangesI'm planning to remove the V2 support. Now, HotColdSchema is one This patch derives: RecordSerializationRoundTripVersion4HotColdSchema from: RecordSerializationRoundTripVersion2HotColdSchema and then adds DummyMap, a V4-speicifc item. Full diff: https://github.com/llvm/llvm-project/pull/156783.diff 1 Files Affected:
diff --git a/llvm/unittests/ProfileData/MemProfTest.cpp b/llvm/unittests/ProfileData/MemProfTest.cpp
index b57567e29f1c8..abe36bc759658 100644
--- a/llvm/unittests/ProfileData/MemProfTest.cpp
+++ b/llvm/unittests/ProfileData/MemProfTest.cpp
@@ -404,6 +404,75 @@ TEST(MemProf, RecordSerializationRoundTripVersion2HotColdSchema) {
EXPECT_EQ(Record, GotRecord);
}
+TEST(MemProf, RecordSerializationRoundTripVersion4HotColdSchema) {
+ const auto Schema = getHotColdSchema();
+
+ MemInfoBlock Info;
+ Info.AllocCount = 11;
+ Info.TotalSize = 22;
+ Info.TotalLifetime = 33;
+ Info.TotalLifetimeAccessDensity = 44;
+
+ llvm::SmallVector<CallStackId> CallStackIds = {0x123, 0x456};
+
+ llvm::SmallVector<CallStackId> CallSiteIds = {0x333, 0x444};
+
+ IndexedMemProfRecord Record;
+ for (const auto &CSId : CallStackIds) {
+ // Use the same info block for both allocation sites.
+ Record.AllocSites.emplace_back(CSId, Info, Schema);
+ }
+ for (auto CSId : CallSiteIds)
+ Record.CallSites.push_back(IndexedCallSiteInfo(CSId));
+
+ std::bitset<llvm::to_underlying(Meta::Size)> SchemaBitSet;
+ for (auto Id : Schema)
+ SchemaBitSet.set(llvm::to_underlying(Id));
+
+ // Verify that SchemaBitSet has the fields we expect and nothing else, which
+ // we check with count().
+ EXPECT_EQ(SchemaBitSet.count(), 4U);
+ EXPECT_TRUE(SchemaBitSet[llvm::to_underlying(Meta::AllocCount)]);
+ EXPECT_TRUE(SchemaBitSet[llvm::to_underlying(Meta::TotalSize)]);
+ EXPECT_TRUE(SchemaBitSet[llvm::to_underlying(Meta::TotalLifetime)]);
+ EXPECT_TRUE(
+ SchemaBitSet[llvm::to_underlying(Meta::TotalLifetimeAccessDensity)]);
+
+ // Verify that Schema has propagated all the way to the Info field in each
+ // IndexedAllocationInfo.
+ ASSERT_THAT(Record.AllocSites, SizeIs(2));
+ EXPECT_EQ(Record.AllocSites[0].Info.getSchema(), SchemaBitSet);
+ EXPECT_EQ(Record.AllocSites[1].Info.getSchema(), SchemaBitSet);
+
+ std::string Buffer;
+ llvm::raw_string_ostream OS(Buffer);
+ // Need a dummy map for V4 serialization
+ llvm::DenseMap<CallStackId, LinearCallStackId> DummyMap = {
+ {0x123, 1}, {0x456, 2}, {0x333, 3}, {0x444, 4}};
+ Record.serialize(Schema, OS, Version4, &DummyMap);
+
+ const IndexedMemProfRecord GotRecord = IndexedMemProfRecord::deserialize(
+ Schema, reinterpret_cast<const unsigned char *>(Buffer.data()), Version4);
+
+ // Verify that Schema comes back correctly after deserialization. Technically,
+ // the comparison between Record and GotRecord below includes the comparison
+ // of their Schemas, but we'll verify the Schemas on our own.
+ ASSERT_THAT(GotRecord.AllocSites, SizeIs(2));
+ EXPECT_EQ(GotRecord.AllocSites[0].Info.getSchema(), SchemaBitSet);
+ EXPECT_EQ(GotRecord.AllocSites[1].Info.getSchema(), SchemaBitSet);
+
+ // Create the expected record using the linear IDs from the dummy map.
+ IndexedMemProfRecord ExpectedRecord;
+ for (const auto &CSId : CallStackIds) {
+ ExpectedRecord.AllocSites.emplace_back(DummyMap[CSId], Info, Schema);
+ }
+ for (const auto &CSId : CallSiteIds) {
+ ExpectedRecord.CallSites.emplace_back(DummyMap[CSId]);
+ }
+
+ EXPECT_EQ(ExpectedRecord, GotRecord);
+}
+
TEST(MemProf, SymbolizationFilter) {
auto Symbolizer = std::make_unique<MockSymbolizer>();
|
teresajohnson
approved these changes
Sep 4, 2025
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/65/builds/22201 Here is the relevant piece of the build log for the reference |
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.
I'm planning to remove the V2 support. Now, HotColdSchema is one
feature that's tested in V2 but not in V4.
This patch derives:
RecordSerializationRoundTripVersion4HotColdSchema
from:
RecordSerializationRoundTripVersion2HotColdSchema
and then adds DummyMap, a V4-speicifc item.