Skip to content

Use an abbrev to reduce size of VALUE_GUID records in ThinLTO summaries #90497

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 1 commit into from
Apr 30, 2024
Merged

Use an abbrev to reduce size of VALUE_GUID records in ThinLTO summaries #90497

merged 1 commit into from
Apr 30, 2024

Conversation

jvoung
Copy link
Contributor

@jvoung jvoung commented Apr 29, 2024

GUID often have content in the higher bits of a 64-bit entry so using
the unabbrev encoding is inefficient (lots of VBR control bits).
Instead, use an abbrev with two 32-bit fixed width chunks.
The abbrev also helps encode the "count" in one place instead of
in every record.

Reduces size of distributed backend summary files by 8.7% in one
example app.

@jvoung
Use an abbrev for VALUE_GUID records in thinlto summaries
f18c255 
The GUID often have content in the higher bits of a 64-bit value
so using the default unabbrev encoding (approx VBR 6) is inefficient,
since many bits are just to encode that there is a following VBR block.

Also, the record always has N fields so the abbrev could help indicate
that N count once.

Reduces size of distributed index files by about 8.7% in one
example app.
@github-actions GitHub Actions
Copy link

Thank you for submitting a Pull Request (PR) to the LLVM Project!

This PR will be automatically labeled and the relevant teams will be
notified.

If you wish to, you can add reviewers by using the "Reviewers" section on this page.

If this is not working for you, it is probably because you do not have write
permissions for the repository. In which case you can instead tag reviewers by
name in a comment by using @ followed by their GitHub username.

If you have received no comments on your PR for a week, you can request a review
by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate
is once a week. Please remember that you are asking for valuable time from other developers.

If you have further questions, they may be answered by the LLVM GitHub User Guide.

You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums.

@jvoung jvoung marked this pull request as ready for review April 29, 2024 17:20
@llvmbot llvmbot added LTO Link time optimization (regular/full LTO or ThinLTO) llvm:ir labels Apr 29, 2024
@llvmbot
Copy link
Member

llvmbot commented Apr 29, 2024

@llvm/pr-subscribers-llvm-ir

Author: Jan Voung (jvoung)

Changes

GUID often have content in the higher bits of a 64-bit entry so using
the unabbrev encoding is inefficient (lots of VBR control bits).
Instead, use an abbrev with two 32-bit fixed width chunks.
The abbrev also helps encode the "count" in one place instead of
in every record.

Reduces size of distributed backend summary files by 8.7% in one
example app.

Patch is 33.10 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/90497.diff

16 Files Affected:

  • (modified) llvm/include/llvm/IR/ModuleSummaryIndex.h (+1-1)
  • (modified) llvm/lib/Bitcode/Reader/BitcodeReader.cpp (+7-2)
  • (modified) llvm/lib/Bitcode/Writer/BitcodeWriter.cpp (+26-4)
  • (modified) llvm/test/Assembler/thinlto-summary.ll (+29-24)
  • (modified) llvm/test/Bitcode/summary_version.ll (+1-1)
  • (modified) llvm/test/Bitcode/thinlto-alias.ll (+2-2)
  • (modified) llvm/test/Bitcode/thinlto-func-summary-vtableref-pgo.ll (+4-4)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-callgraph-partial-sample-profile-summary.ll (+1-1)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-callgraph-pgo.ll (+1-1)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-callgraph-profile-summary.ll (+1-1)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-callgraph-sample-profile-summary.ll (+1-1)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-callgraph.ll (+1-1)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-originalnames.ll (+3-3)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-paramaccess.ll (+38-38)
  • (modified) llvm/test/ThinLTO/X86/distributed_indexes.ll (+8-8)
  • (modified) llvm/test/tools/llvm-lto/thinlto.ll (+2-2)
diff --git a/llvm/include/llvm/IR/ModuleSummaryIndex.h b/llvm/include/llvm/IR/ModuleSummaryIndex.h
index 5d137d4b3553cf..fa2a7b42c9aab3 100644
--- a/llvm/include/llvm/IR/ModuleSummaryIndex.h
+++ b/llvm/include/llvm/IR/ModuleSummaryIndex.h
@@ -1423,7 +1423,7 @@ class ModuleSummaryIndex {
   // in the way some record are interpreted, like flags for instance.
   // Note that incrementing this may require changes in both BitcodeReader.cpp
   // and BitcodeWriter.cpp.
-  static constexpr uint64_t BitcodeSummaryVersion = 9;
+  static constexpr uint64_t BitcodeSummaryVersion = 10;
 
   // Regular LTO module name for ASM writer
   static constexpr const char *getRegularLTOModuleName() {
diff --git a/llvm/lib/Bitcode/Reader/BitcodeReader.cpp b/llvm/lib/Bitcode/Reader/BitcodeReader.cpp
index 0b7fcd88418894..a0779f955cf28d 100644
--- a/llvm/lib/Bitcode/Reader/BitcodeReader.cpp
+++ b/llvm/lib/Bitcode/Reader/BitcodeReader.cpp
@@ -7513,9 +7513,14 @@ Error ModuleSummaryIndexBitcodeReader::parseEntireSummary(unsigned ID) {
       TheIndex.setFlags(Record[0]);
       break;
     }
-    case bitc::FS_VALUE_GUID: { // [valueid, refguid]
+    case bitc::FS_VALUE_GUID: { // [valueid, refguid_upper32, refguid_lower32]
       uint64_t ValueID = Record[0];
-      GlobalValue::GUID RefGUID = Record[1];
+      GlobalValue::GUID RefGUID;
+      if (Version >= 10) {
+        RefGUID = Record[1] << 32 | Record[2];
+      } else {
+        RefGUID = Record[1];
+      }
       ValueIdToValueInfoMap[ValueID] = std::make_tuple(
           TheIndex.getOrInsertValueInfo(RefGUID), RefGUID, RefGUID);
       break;
diff --git a/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp b/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp
index 6d01e3b4d82189..1aaf160e91ca18 100644
--- a/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp
+++ b/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp
@@ -4299,9 +4299,20 @@ void ModuleBitcodeWriterBase::writePerModuleGlobalValueSummary() {
     return;
   }
 
+  auto Abbv = std::make_shared<BitCodeAbbrev>();
+  Abbv->Add(BitCodeAbbrevOp(bitc::FS_VALUE_GUID));
+  Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 6));
+  // GUIDS often use up most of 64-bits, so encode as two Fixed 32.
+  Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 32));
+  Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 32));
+  unsigned ValueGuidAbbrev = Stream.EmitAbbrev(std::move(Abbv));
+
   for (const auto &GVI : valueIds()) {
     Stream.EmitRecord(bitc::FS_VALUE_GUID,
-                      ArrayRef<uint64_t>{GVI.second, GVI.first});
+                      ArrayRef<uint32_t>{GVI.second,
+                                         static_cast<uint32_t>(GVI.first >> 32),
+                                         static_cast<uint32_t>(GVI.first)},
+                      ValueGuidAbbrev);
   }
 
   if (!Index->stackIds().empty()) {
@@ -4315,7 +4326,7 @@ void ModuleBitcodeWriterBase::writePerModuleGlobalValueSummary() {
   }
 
   // Abbrev for FS_PERMODULE_PROFILE.
-  auto Abbv = std::make_shared<BitCodeAbbrev>();
+  Abbv = std::make_shared<BitCodeAbbrev>();
   Abbv->Add(BitCodeAbbrevOp(bitc::FS_PERMODULE_PROFILE));
   Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8));   // valueid
   Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8));   // flags
@@ -4467,9 +4478,20 @@ void IndexBitcodeWriter::writeCombinedGlobalValueSummary() {
   // Write the index flags.
   Stream.EmitRecord(bitc::FS_FLAGS, ArrayRef<uint64_t>{Index.getFlags()});
 
+  auto Abbv = std::make_shared<BitCodeAbbrev>();
+  Abbv->Add(BitCodeAbbrevOp(bitc::FS_VALUE_GUID));
+  Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 6));
+  // GUIDS often use up most of 64-bits, so encode as two Fixed 32.
+  Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 32));
+  Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 32));
+  unsigned ValueGuidAbbrev = Stream.EmitAbbrev(std::move(Abbv));
+
   for (const auto &GVI : valueIds()) {
     Stream.EmitRecord(bitc::FS_VALUE_GUID,
-                      ArrayRef<uint64_t>{GVI.second, GVI.first});
+                      ArrayRef<uint32_t>{GVI.second,
+                                         static_cast<uint32_t>(GVI.first >> 32),
+                                         static_cast<uint32_t>(GVI.first)},
+                      ValueGuidAbbrev);
   }
 
   if (!StackIdIndices.empty()) {
@@ -4488,7 +4510,7 @@ void IndexBitcodeWriter::writeCombinedGlobalValueSummary() {
   }
 
   // Abbrev for FS_COMBINED_PROFILE.
-  auto Abbv = std::make_shared<BitCodeAbbrev>();
+  Abbv = std::make_shared<BitCodeAbbrev>();
   Abbv->Add(BitCodeAbbrevOp(bitc::FS_COMBINED_PROFILE));
   Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8));   // valueid
   Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8));   // modid
diff --git a/llvm/test/Assembler/thinlto-summary.ll b/llvm/test/Assembler/thinlto-summary.ll
index 05dad2c7acad46..e0d866da0d8a22 100644
--- a/llvm/test/Assembler/thinlto-summary.ll
+++ b/llvm/test/Assembler/thinlto-summary.ll
@@ -46,28 +46,32 @@
 ^18 = gv: (guid: 17, summaries: (alias: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 1), aliasee: ^14)))
 
 ; Test all types of TypeIdInfo on function summaries.
-^19 = gv: (guid: 18, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 4, typeIdInfo: (typeTests: (^25, ^27)))))
-^20 = gv: (guid: 19, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 8, typeIdInfo: (typeTestAssumeVCalls: (vFuncId: (^28, offset: 16))))))
-^21 = gv: (guid: 20, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 5, typeIdInfo: (typeCheckedLoadVCalls: (vFuncId: (^26, offset: 16))))))
-^22 = gv: (guid: 21, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 15, typeIdInfo: (typeTestAssumeConstVCalls: ((vFuncId: (^28, offset: 16), args: (42)), (vFuncId: (^28, offset: 24)))))))
-^23 = gv: (guid: 22, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 5, typeIdInfo: (typeCheckedLoadConstVCalls: ((vFuncId: (^29, offset: 16), args: (42)))))))
+^19 = gv: (guid: 18, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 4, typeIdInfo: (typeTests: (^26, ^28)))))
+^20 = gv: (guid: 19, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 8, typeIdInfo: (typeTestAssumeVCalls: (vFuncId: (^29, offset: 16))))))
+^21 = gv: (guid: 20, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 5, typeIdInfo: (typeCheckedLoadVCalls: (vFuncId: (^27, offset: 16))))))
+^22 = gv: (guid: 21, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 15, typeIdInfo: (typeTestAssumeConstVCalls: ((vFuncId: (^29, offset: 16), args: (42)), (vFuncId: (^29, offset: 24)))))))
+^23 = gv: (guid: 22, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 5, typeIdInfo: (typeCheckedLoadConstVCalls: ((vFuncId: (^30, offset: 16), args: (42)))))))
 
 ; Function summary with an import type of declaration
 ^24 = gv: (guid: 23, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, importType: declaration), insts: 5)))
 
+; GUID that are 64-bit
+
+^25 = gv: (guid: 9123456789101112131, summaries: (function: (module: ^0, flags: (linkage: internal, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 1, importType: definition), insts: 1)))
+
 ; Test TypeId summaries:
 
-^25 = typeid: (name: "_ZTS1C", summary: (typeTestRes: (kind: single, sizeM1BitWidth: 0)))
+^26 = typeid: (name: "_ZTS1C", summary: (typeTestRes: (kind: single, sizeM1BitWidth: 0)))
 ; Test TypeId with other optional fields (alignLog2/sizeM1/bitMask/inlineBits)
-^26 = typeid: (name: "_ZTS1B", summary: (typeTestRes: (kind: inline, sizeM1BitWidth: 0, alignLog2: 1, sizeM1: 2, bitMask: 3, inlineBits: 4)))
+^27 = typeid: (name: "_ZTS1B", summary: (typeTestRes: (kind: inline, sizeM1BitWidth: 0, alignLog2: 1, sizeM1: 2, bitMask: 3, inlineBits: 4)))
 ; Test the AllOnes resolution, and all kinds of WholeProgramDevirtResolution
 ; types, including all optional resolution by argument kinds.
-^27 = typeid: (name: "_ZTS1A", summary: (typeTestRes: (kind: allOnes, sizeM1BitWidth: 7), wpdResolutions: ((offset: 0, wpdRes: (kind: branchFunnel)), (offset: 8, wpdRes: (kind: singleImpl, singleImplName: "_ZN1A1nEi")), (offset: 16, wpdRes: (kind: indir, resByArg: (args: (1, 2), byArg: (kind: indir, byte: 2, bit: 3), args: (3), byArg: (kind: uniformRetVal, info: 1), args: (4), byArg: (kind: uniqueRetVal, info: 1), args: (5), byArg: (kind: virtualConstProp)))))))
+^28 = typeid: (name: "_ZTS1A", summary: (typeTestRes: (kind: allOnes, sizeM1BitWidth: 7), wpdResolutions: ((offset: 0, wpdRes: (kind: branchFunnel)), (offset: 8, wpdRes: (kind: singleImpl, singleImplName: "_ZN1A1nEi")), (offset: 16, wpdRes: (kind: indir, resByArg: (args: (1, 2), byArg: (kind: indir, byte: 2, bit: 3), args: (3), byArg: (kind: uniformRetVal, info: 1), args: (4), byArg: (kind: uniqueRetVal, info: 1), args: (5), byArg: (kind: virtualConstProp)))))))
 ; Test the other kinds of type test resoultions
-^28 = typeid: (name: "_ZTS1D", summary: (typeTestRes: (kind: byteArray, sizeM1BitWidth: 0)))
-^29 = typeid: (name: "_ZTS1E", summary: (typeTestRes: (kind: unsat, sizeM1BitWidth: 0)))
-^30 = flags: 8
-^31 = blockcount: 1888
+^29 = typeid: (name: "_ZTS1D", summary: (typeTestRes: (kind: byteArray, sizeM1BitWidth: 0)))
+^30 = typeid: (name: "_ZTS1E", summary: (typeTestRes: (kind: unsat, sizeM1BitWidth: 0)))
+^31 = flags: 8
+^32 = blockcount: 1888
 
 ; Make sure we get back from llvm-dis essentially what we put in via llvm-as.
 ; CHECK: ^0 = module: (path: "thinlto-summary1.o", hash: (1369602428, 2747878711, 259090915, 2507395659, 1141468049))
@@ -91,19 +95,20 @@
 ; CHECK: ^16 = gv: (guid: 15, summaries: (function: (module: ^1, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 1, funcFlags: (readNone: 1, readOnly: 0, noRecurse: 1, returnDoesNotAlias: 0, noInline: 0, alwaysInline: 1, noUnwind: 1, mayThrow: 1, hasUnknownCall: 1, mustBeUnreachable: 0))))
 ; CHECK: ^17 = gv: (guid: 16, summaries: (function: (module: ^1, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 1, funcFlags: (readNone: 0, readOnly: 1, noRecurse: 0, returnDoesNotAlias: 1, noInline: 0, alwaysInline: 0, noUnwind: 0, mayThrow: 0, hasUnknownCall: 0, mustBeUnreachable: 1), calls: ((callee: ^15)))))
 ; CHECK: ^18 = gv: (guid: 17, summaries: (alias: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 1, canAutoHide: 0, importType: definition), aliasee: ^14)))
-; CHECK: ^19 = gv: (guid: 18, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 4, typeIdInfo: (typeTests: (^25, ^27)))))
-; CHECK: ^20 = gv: (guid: 19, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 8, typeIdInfo: (typeTestAssumeVCalls: (vFuncId: (^28, offset: 16))))))
-; CHECK: ^21 = gv: (guid: 20, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 5, typeIdInfo: (typeCheckedLoadVCalls: (vFuncId: (^26, offset: 16))))))
-; CHECK: ^22 = gv: (guid: 21, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 15, typeIdInfo: (typeTestAssumeConstVCalls: ((vFuncId: (^28, offset: 16), args: (42)), (vFuncId: (^28, offset: 24)))))))
-; CHECK: ^23 = gv: (guid: 22, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 5, typeIdInfo: (typeCheckedLoadConstVCalls: ((vFuncId: (^29, offset: 16), args: (42)))))))
+; CHECK: ^19 = gv: (guid: 18, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 4, typeIdInfo: (typeTests: (^26, ^28)))))
+; CHECK: ^20 = gv: (guid: 19, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 8, typeIdInfo: (typeTestAssumeVCalls: (vFuncId: (^29, offset: 16))))))
+; CHECK: ^21 = gv: (guid: 20, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 5, typeIdInfo: (typeCheckedLoadVCalls: (vFuncId: (^27, offset: 16))))))
+; CHECK: ^22 = gv: (guid: 21, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 15, typeIdInfo: (typeTestAssumeConstVCalls: ((vFuncId: (^29, offset: 16), args: (42)), (vFuncId: (^29, offset: 24)))))))
+; CHECK: ^23 = gv: (guid: 22, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 5, typeIdInfo: (typeCheckedLoadConstVCalls: ((vFuncId: (^30, offset: 16), args: (42)))))))
 ; CHECK: ^24 = gv: (guid: 23, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: declaration), insts: 5)))
-; CHECK: ^25 = typeid: (name: "_ZTS1C", summary: (typeTestRes: (kind: single, sizeM1BitWidth: 0))) ; guid = 1884921850105019584
-; CHECK: ^26 = typeid: (name: "_ZTS1B", summary: (typeTestRes: (kind: inline, sizeM1BitWidth: 0, alignLog2: 1, sizeM1: 2, bitMask: 3, inlineBits: 4))) ; guid = 6203814149063363976
-; CHECK: ^27 = typeid: (name: "_ZTS1A", summary: (typeTestRes: (kind: allOnes, sizeM1BitWidth: 7), wpdResolutions: ((offset: 0, wpdRes: (kind: branchFunnel)), (offset: 8, wpdRes: (kind: singleImpl, singleImplName: "_ZN1A1nEi")), (offset: 16, wpdRes: (kind: indir, resByArg: (args: (1, 2), byArg: (kind: indir, byte: 2, bit: 3), args: (3), byArg: (kind: uniformRetVal, info: 1), args: (4), byArg: (kind: uniqueRetVal, info: 1), args: (5), byArg: (kind: virtualConstProp))))))) ; guid = 7004155349499253778
-; CHECK: ^28 = typeid: (name: "_ZTS1D", summary: (typeTestRes: (kind: byteArray, sizeM1BitWidth: 0))) ; guid = 9614786172484273522
-; CHECK: ^29 = typeid: (name: "_ZTS1E", summary: (typeTestRes: (kind: unsat, sizeM1BitWidth: 0))) ; guid = 17437243864166745132
-; CHECK: ^30 = flags: 8
-; CHECK: ^31 = blockcount: 1888
+; CHECK: ^25 = gv: (guid: 9123456789101112131, summaries: (function: (module: ^0, flags: (linkage: internal, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 1, canAutoHide: 0, importType: definition), insts: 1)))
+; CHECK: ^26 = typeid: (name: "_ZTS1C", summary: (typeTestRes: (kind: single, sizeM1BitWidth: 0))) ; guid = 1884921850105019584
+; CHECK: ^27 = typeid: (name: "_ZTS1B", summary: (typeTestRes: (kind: inline, sizeM1BitWidth: 0, alignLog2: 1, sizeM1: 2, bitMask: 3, inlineBits: 4))) ; guid = 6203814149063363976
+; CHECK: ^28 = typeid: (name: "_ZTS1A", summary: (typeTestRes: (kind: allOnes, sizeM1BitWidth: 7), wpdResolutions: ((offset: 0, wpdRes: (kind: branchFunnel)), (offset: 8, wpdRes: (kind: singleImpl, singleImplName: "_ZN1A1nEi")), (offset: 16, wpdRes: (kind: indir, resByArg: (args: (1, 2), byArg: (kind: indir, byte: 2, bit: 3), args: (3), byArg: (kind: uniformRetVal, info: 1), args: (4), byArg: (kind: uniqueRetVal, info: 1), args: (5), byArg: (kind: virtualConstProp))))))) ; guid = 7004155349499253778
+; CHECK: ^29 = typeid: (name: "_ZTS1D", summary: (typeTestRes: (kind: byteArray, sizeM1BitWidth: 0))) ; guid = 9614786172484273522
+; CHECK: ^30 = typeid: (name: "_ZTS1E", summary: (typeTestRes: (kind: unsat, sizeM1BitWidth: 0))) ; guid = 17437243864166745132
+; CHECK: ^31 = flags: 8
+; CHECK: ^32 = blockcount: 1888
 
 ; Make sure parsing of a non-summary entry containing a ":" does not fail
 ; after summary parsing, which handles colons differently.
diff --git a/llvm/test/Bitcode/summary_version.ll b/llvm/test/Bitcode/summary_version.ll
index 98feab6fe2f995..26c64f81a773f1 100644
--- a/llvm/test/Bitcode/summary_version.ll
+++ b/llvm/test/Bitcode/summary_version.ll
@@ -2,7 +2,7 @@
 ; RUN: opt  -module-summary  %s -o - | llvm-bcanalyzer -dump | FileCheck %s
 
 ; CHECK: <GLOBALVAL_SUMMARY_BLOCK
-; CHECK: <VERSION op0=9/>
+; CHECK: <VERSION op0=10/>
 
 
 
diff --git a/llvm/test/Bitcode/thinlto-alias.ll b/llvm/test/Bitcode/thinlto-alias.ll
index 5dfff0f796198d..7deb2d8259e328 100644
--- a/llvm/test/Bitcode/thinlto-alias.ll
+++ b/llvm/test/Bitcode/thinlto-alias.ll
@@ -31,9 +31,9 @@
 ; COMBINED-NEXT:    <VERSION
 ; COMBINED-NEXT:    <FLAGS
 ; See if the call to analias is registered, using the expected value id.
-; COMBINED-NEXT:    <VALUE_GUID op0=[[ALIASID:[0-9]+]] op1=-5751648690987223394/>
+; COMBINED-NEXT:    <VALUE_GUID {{.*}} op0=[[ALIASID:[0-9]+]] op1=2955807229 op2=886945438/>
 ; COMBINED-NEXT:    <VALUE_GUID
-; COMBINED-NEXT:    <VALUE_GUID op0=[[ALIASEEID:[0-9]+]] op1=-1039159065113703048/>
+; COMBINED-NEXT:    <VALUE_GUID {{.*}} op0=[[ALIASEEID:[0-9]+]] op1=4053019222 op2=46484856/>
 ; COMBINED-NEXT:    <COMBINED_PROFILE {{.*}} op9=[[ALIASID]]
 ; COMBINED-NEXT:    <COMBINED_PROFILE {{.*}}
 ; COMBINED-NEXT:    <COMBINED_ALIAS  {{.*}} op3=[[ALIASEEID]]
diff --git a/llvm/test/Bitcode/thinlto-func-summary-vtableref-pgo.ll b/llvm/test/Bitcode/thinlto-func-summary-vtableref-pgo.ll
index 19e228fd5355c7..d864cadcff3541 100644
--- a/llvm/test/Bitcode/thinlto-func-summary-vtableref-pgo.ll
+++ b/llvm/test/Bitcode/thinlto-func-summary-vtableref-pgo.ll
@@ -11,21 +11,21 @@
 ; RUN: llvm-dis -o - %t.o | llvm-as -o - | llvm-dis -o - | FileCheck %s --check-prefix=DIS
 
 ; CHECK: <GLOBALVAL_SUMMARY_BLOCK
-; CHECK-NEXT:   <VERSION op0=9/>
+; CHECK-NEXT:   <VERSION op0=10/>
 ; CHECK-NEXT:   <FLAGS op0=0/>
 ; The `VALUE_GUID` below represents the "_ZTV4Base" referenced by the instruction
 ; that loads vtable pointers.
-; CHECK-NEXT: <VALUE_GUID op0=21 op1=1960855528937986108/>
+; CHECK-NEXT: <VALUE_GUID {{.*}} op0=21 op1=456547254 op2=3929380924/>
 ; The `VALUE_GUID` below represents the "_ZN4Base4funcEv" referenced by the
 ; indirect call instruction.
-; CHECK-NEXT:      <VALUE_GUID op0=20 op1=5459407273543877811/>
+; CHECK-NEXT:      <VALUE_GUID {{.*}} op0=20 op1=1271117309 op2=2009351347/>
 ; NOTE vtables and functions from Derived class is dropped because
 ; `-icp-max-num-vtables` and `-icp-max-prom` are both set to one.
 ; <PERMODULE_PROFILE> has the format [valueid, flags, instcount, funcflags,
 ;                                     numrefs, rorefcnt, worefcnt,
 ;                                     m x valueid,
 ;                                     n x (valueid, hotness+tailcall)]
-; CHECK-NEXT:   <PERMODULE_PROFILE abbrevid=4 op0=0 op1=0 op2=4 op3=256 op4=1 op5=1 op6=0 op7=21 op8=20 op9=3/>
+; CHECK-NEXT:   <PERMODULE_PROFILE {{.*}} op0=0 op1=0 op2=4 op3=256 op4=1 op5=1 op6=0 op7=21 op8=20 op9=3/>
 ; CHECK-NEXT:  </GLOBALVAL_SUMMARY_BLOCK>
 
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
diff --git a/llvm/test/Bitcode/thinlto-function-summary-callgraph-partial-sample-profile-summary.ll b/llvm/test/Bitcode/thinlto-function-summary-callgraph-partial-sample-profile-summary.ll
index d44ee24694be56..0c3ab9b2089310 10...
[truncated]

@llvmbot
Copy link
Member

llvmbot commented Apr 29, 2024

@llvm/pr-subscribers-lto

Author: Jan Voung (jvoung)

Changes

GUID often have content in the higher bits of a 64-bit entry so using
the unabbrev encoding is inefficient (lots of VBR control bits).
Instead, use an abbrev with two 32-bit fixed width chunks.
The abbrev also helps encode the "count" in one place instead of
in every record.

Reduces size of distributed backend summary files by 8.7% in one
example app.

Patch is 33.10 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/90497.diff

16 Files Affected:

  • (modified) llvm/include/llvm/IR/ModuleSummaryIndex.h (+1-1)
  • (modified) llvm/lib/Bitcode/Reader/BitcodeReader.cpp (+7-2)
  • (modified) llvm/lib/Bitcode/Writer/BitcodeWriter.cpp (+26-4)
  • (modified) llvm/test/Assembler/thinlto-summary.ll (+29-24)
  • (modified) llvm/test/Bitcode/summary_version.ll (+1-1)
  • (modified) llvm/test/Bitcode/thinlto-alias.ll (+2-2)
  • (modified) llvm/test/Bitcode/thinlto-func-summary-vtableref-pgo.ll (+4-4)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-callgraph-partial-sample-profile-summary.ll (+1-1)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-callgraph-pgo.ll (+1-1)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-callgraph-profile-summary.ll (+1-1)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-callgraph-sample-profile-summary.ll (+1-1)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-callgraph.ll (+1-1)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-originalnames.ll (+3-3)
  • (modified) llvm/test/Bitcode/thinlto-function-summary-paramaccess.ll (+38-38)
  • (modified) llvm/test/ThinLTO/X86/distributed_indexes.ll (+8-8)
  • (modified) llvm/test/tools/llvm-lto/thinlto.ll (+2-2)
diff --git a/llvm/include/llvm/IR/ModuleSummaryIndex.h b/llvm/include/llvm/IR/ModuleSummaryIndex.h
index 5d137d4b3553cf..fa2a7b42c9aab3 100644
--- a/llvm/include/llvm/IR/ModuleSummaryIndex.h
+++ b/llvm/include/llvm/IR/ModuleSummaryIndex.h
@@ -1423,7 +1423,7 @@ class ModuleSummaryIndex {
   // in the way some record are interpreted, like flags for instance.
   // Note that incrementing this may require changes in both BitcodeReader.cpp
   // and BitcodeWriter.cpp.
-  static constexpr uint64_t BitcodeSummaryVersion = 9;
+  static constexpr uint64_t BitcodeSummaryVersion = 10;
 
   // Regular LTO module name for ASM writer
   static constexpr const char *getRegularLTOModuleName() {
diff --git a/llvm/lib/Bitcode/Reader/BitcodeReader.cpp b/llvm/lib/Bitcode/Reader/BitcodeReader.cpp
index 0b7fcd88418894..a0779f955cf28d 100644
--- a/llvm/lib/Bitcode/Reader/BitcodeReader.cpp
+++ b/llvm/lib/Bitcode/Reader/BitcodeReader.cpp
@@ -7513,9 +7513,14 @@ Error ModuleSummaryIndexBitcodeReader::parseEntireSummary(unsigned ID) {
       TheIndex.setFlags(Record[0]);
       break;
     }
-    case bitc::FS_VALUE_GUID: { // [valueid, refguid]
+    case bitc::FS_VALUE_GUID: { // [valueid, refguid_upper32, refguid_lower32]
       uint64_t ValueID = Record[0];
-      GlobalValue::GUID RefGUID = Record[1];
+      GlobalValue::GUID RefGUID;
+      if (Version >= 10) {
+        RefGUID = Record[1] << 32 | Record[2];
+      } else {
+        RefGUID = Record[1];
+      }
       ValueIdToValueInfoMap[ValueID] = std::make_tuple(
           TheIndex.getOrInsertValueInfo(RefGUID), RefGUID, RefGUID);
       break;
diff --git a/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp b/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp
index 6d01e3b4d82189..1aaf160e91ca18 100644
--- a/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp
+++ b/llvm/lib/Bitcode/Writer/BitcodeWriter.cpp
@@ -4299,9 +4299,20 @@ void ModuleBitcodeWriterBase::writePerModuleGlobalValueSummary() {
     return;
   }
 
+  auto Abbv = std::make_shared<BitCodeAbbrev>();
+  Abbv->Add(BitCodeAbbrevOp(bitc::FS_VALUE_GUID));
+  Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 6));
+  // GUIDS often use up most of 64-bits, so encode as two Fixed 32.
+  Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 32));
+  Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 32));
+  unsigned ValueGuidAbbrev = Stream.EmitAbbrev(std::move(Abbv));
+
   for (const auto &GVI : valueIds()) {
     Stream.EmitRecord(bitc::FS_VALUE_GUID,
-                      ArrayRef<uint64_t>{GVI.second, GVI.first});
+                      ArrayRef<uint32_t>{GVI.second,
+                                         static_cast<uint32_t>(GVI.first >> 32),
+                                         static_cast<uint32_t>(GVI.first)},
+                      ValueGuidAbbrev);
   }
 
   if (!Index->stackIds().empty()) {
@@ -4315,7 +4326,7 @@ void ModuleBitcodeWriterBase::writePerModuleGlobalValueSummary() {
   }
 
   // Abbrev for FS_PERMODULE_PROFILE.
-  auto Abbv = std::make_shared<BitCodeAbbrev>();
+  Abbv = std::make_shared<BitCodeAbbrev>();
   Abbv->Add(BitCodeAbbrevOp(bitc::FS_PERMODULE_PROFILE));
   Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8));   // valueid
   Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8));   // flags
@@ -4467,9 +4478,20 @@ void IndexBitcodeWriter::writeCombinedGlobalValueSummary() {
   // Write the index flags.
   Stream.EmitRecord(bitc::FS_FLAGS, ArrayRef<uint64_t>{Index.getFlags()});
 
+  auto Abbv = std::make_shared<BitCodeAbbrev>();
+  Abbv->Add(BitCodeAbbrevOp(bitc::FS_VALUE_GUID));
+  Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 6));
+  // GUIDS often use up most of 64-bits, so encode as two Fixed 32.
+  Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 32));
+  Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::Fixed, 32));
+  unsigned ValueGuidAbbrev = Stream.EmitAbbrev(std::move(Abbv));
+
   for (const auto &GVI : valueIds()) {
     Stream.EmitRecord(bitc::FS_VALUE_GUID,
-                      ArrayRef<uint64_t>{GVI.second, GVI.first});
+                      ArrayRef<uint32_t>{GVI.second,
+                                         static_cast<uint32_t>(GVI.first >> 32),
+                                         static_cast<uint32_t>(GVI.first)},
+                      ValueGuidAbbrev);
   }
 
   if (!StackIdIndices.empty()) {
@@ -4488,7 +4510,7 @@ void IndexBitcodeWriter::writeCombinedGlobalValueSummary() {
   }
 
   // Abbrev for FS_COMBINED_PROFILE.
-  auto Abbv = std::make_shared<BitCodeAbbrev>();
+  Abbv = std::make_shared<BitCodeAbbrev>();
   Abbv->Add(BitCodeAbbrevOp(bitc::FS_COMBINED_PROFILE));
   Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8));   // valueid
   Abbv->Add(BitCodeAbbrevOp(BitCodeAbbrevOp::VBR, 8));   // modid
diff --git a/llvm/test/Assembler/thinlto-summary.ll b/llvm/test/Assembler/thinlto-summary.ll
index 05dad2c7acad46..e0d866da0d8a22 100644
--- a/llvm/test/Assembler/thinlto-summary.ll
+++ b/llvm/test/Assembler/thinlto-summary.ll
@@ -46,28 +46,32 @@
 ^18 = gv: (guid: 17, summaries: (alias: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 1), aliasee: ^14)))
 
 ; Test all types of TypeIdInfo on function summaries.
-^19 = gv: (guid: 18, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 4, typeIdInfo: (typeTests: (^25, ^27)))))
-^20 = gv: (guid: 19, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 8, typeIdInfo: (typeTestAssumeVCalls: (vFuncId: (^28, offset: 16))))))
-^21 = gv: (guid: 20, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 5, typeIdInfo: (typeCheckedLoadVCalls: (vFuncId: (^26, offset: 16))))))
-^22 = gv: (guid: 21, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 15, typeIdInfo: (typeTestAssumeConstVCalls: ((vFuncId: (^28, offset: 16), args: (42)), (vFuncId: (^28, offset: 24)))))))
-^23 = gv: (guid: 22, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 5, typeIdInfo: (typeCheckedLoadConstVCalls: ((vFuncId: (^29, offset: 16), args: (42)))))))
+^19 = gv: (guid: 18, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 4, typeIdInfo: (typeTests: (^26, ^28)))))
+^20 = gv: (guid: 19, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 8, typeIdInfo: (typeTestAssumeVCalls: (vFuncId: (^29, offset: 16))))))
+^21 = gv: (guid: 20, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 5, typeIdInfo: (typeCheckedLoadVCalls: (vFuncId: (^27, offset: 16))))))
+^22 = gv: (guid: 21, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 15, typeIdInfo: (typeTestAssumeConstVCalls: ((vFuncId: (^29, offset: 16), args: (42)), (vFuncId: (^29, offset: 24)))))))
+^23 = gv: (guid: 22, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0), insts: 5, typeIdInfo: (typeCheckedLoadConstVCalls: ((vFuncId: (^30, offset: 16), args: (42)))))))
 
 ; Function summary with an import type of declaration
 ^24 = gv: (guid: 23, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, importType: declaration), insts: 5)))
 
+; GUID that are 64-bit
+
+^25 = gv: (guid: 9123456789101112131, summaries: (function: (module: ^0, flags: (linkage: internal, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 1, importType: definition), insts: 1)))
+
 ; Test TypeId summaries:
 
-^25 = typeid: (name: "_ZTS1C", summary: (typeTestRes: (kind: single, sizeM1BitWidth: 0)))
+^26 = typeid: (name: "_ZTS1C", summary: (typeTestRes: (kind: single, sizeM1BitWidth: 0)))
 ; Test TypeId with other optional fields (alignLog2/sizeM1/bitMask/inlineBits)
-^26 = typeid: (name: "_ZTS1B", summary: (typeTestRes: (kind: inline, sizeM1BitWidth: 0, alignLog2: 1, sizeM1: 2, bitMask: 3, inlineBits: 4)))
+^27 = typeid: (name: "_ZTS1B", summary: (typeTestRes: (kind: inline, sizeM1BitWidth: 0, alignLog2: 1, sizeM1: 2, bitMask: 3, inlineBits: 4)))
 ; Test the AllOnes resolution, and all kinds of WholeProgramDevirtResolution
 ; types, including all optional resolution by argument kinds.
-^27 = typeid: (name: "_ZTS1A", summary: (typeTestRes: (kind: allOnes, sizeM1BitWidth: 7), wpdResolutions: ((offset: 0, wpdRes: (kind: branchFunnel)), (offset: 8, wpdRes: (kind: singleImpl, singleImplName: "_ZN1A1nEi")), (offset: 16, wpdRes: (kind: indir, resByArg: (args: (1, 2), byArg: (kind: indir, byte: 2, bit: 3), args: (3), byArg: (kind: uniformRetVal, info: 1), args: (4), byArg: (kind: uniqueRetVal, info: 1), args: (5), byArg: (kind: virtualConstProp)))))))
+^28 = typeid: (name: "_ZTS1A", summary: (typeTestRes: (kind: allOnes, sizeM1BitWidth: 7), wpdResolutions: ((offset: 0, wpdRes: (kind: branchFunnel)), (offset: 8, wpdRes: (kind: singleImpl, singleImplName: "_ZN1A1nEi")), (offset: 16, wpdRes: (kind: indir, resByArg: (args: (1, 2), byArg: (kind: indir, byte: 2, bit: 3), args: (3), byArg: (kind: uniformRetVal, info: 1), args: (4), byArg: (kind: uniqueRetVal, info: 1), args: (5), byArg: (kind: virtualConstProp)))))))
 ; Test the other kinds of type test resoultions
-^28 = typeid: (name: "_ZTS1D", summary: (typeTestRes: (kind: byteArray, sizeM1BitWidth: 0)))
-^29 = typeid: (name: "_ZTS1E", summary: (typeTestRes: (kind: unsat, sizeM1BitWidth: 0)))
-^30 = flags: 8
-^31 = blockcount: 1888
+^29 = typeid: (name: "_ZTS1D", summary: (typeTestRes: (kind: byteArray, sizeM1BitWidth: 0)))
+^30 = typeid: (name: "_ZTS1E", summary: (typeTestRes: (kind: unsat, sizeM1BitWidth: 0)))
+^31 = flags: 8
+^32 = blockcount: 1888
 
 ; Make sure we get back from llvm-dis essentially what we put in via llvm-as.
 ; CHECK: ^0 = module: (path: "thinlto-summary1.o", hash: (1369602428, 2747878711, 259090915, 2507395659, 1141468049))
@@ -91,19 +95,20 @@
 ; CHECK: ^16 = gv: (guid: 15, summaries: (function: (module: ^1, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 1, funcFlags: (readNone: 1, readOnly: 0, noRecurse: 1, returnDoesNotAlias: 0, noInline: 0, alwaysInline: 1, noUnwind: 1, mayThrow: 1, hasUnknownCall: 1, mustBeUnreachable: 0))))
 ; CHECK: ^17 = gv: (guid: 16, summaries: (function: (module: ^1, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 1, funcFlags: (readNone: 0, readOnly: 1, noRecurse: 0, returnDoesNotAlias: 1, noInline: 0, alwaysInline: 0, noUnwind: 0, mayThrow: 0, hasUnknownCall: 0, mustBeUnreachable: 1), calls: ((callee: ^15)))))
 ; CHECK: ^18 = gv: (guid: 17, summaries: (alias: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 1, canAutoHide: 0, importType: definition), aliasee: ^14)))
-; CHECK: ^19 = gv: (guid: 18, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 4, typeIdInfo: (typeTests: (^25, ^27)))))
-; CHECK: ^20 = gv: (guid: 19, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 8, typeIdInfo: (typeTestAssumeVCalls: (vFuncId: (^28, offset: 16))))))
-; CHECK: ^21 = gv: (guid: 20, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 5, typeIdInfo: (typeCheckedLoadVCalls: (vFuncId: (^26, offset: 16))))))
-; CHECK: ^22 = gv: (guid: 21, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 15, typeIdInfo: (typeTestAssumeConstVCalls: ((vFuncId: (^28, offset: 16), args: (42)), (vFuncId: (^28, offset: 24)))))))
-; CHECK: ^23 = gv: (guid: 22, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 5, typeIdInfo: (typeCheckedLoadConstVCalls: ((vFuncId: (^29, offset: 16), args: (42)))))))
+; CHECK: ^19 = gv: (guid: 18, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 4, typeIdInfo: (typeTests: (^26, ^28)))))
+; CHECK: ^20 = gv: (guid: 19, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 8, typeIdInfo: (typeTestAssumeVCalls: (vFuncId: (^29, offset: 16))))))
+; CHECK: ^21 = gv: (guid: 20, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 5, typeIdInfo: (typeCheckedLoadVCalls: (vFuncId: (^27, offset: 16))))))
+; CHECK: ^22 = gv: (guid: 21, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 15, typeIdInfo: (typeTestAssumeConstVCalls: ((vFuncId: (^29, offset: 16), args: (42)), (vFuncId: (^29, offset: 24)))))))
+; CHECK: ^23 = gv: (guid: 22, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: definition), insts: 5, typeIdInfo: (typeCheckedLoadConstVCalls: ((vFuncId: (^30, offset: 16), args: (42)))))))
 ; CHECK: ^24 = gv: (guid: 23, summaries: (function: (module: ^0, flags: (linkage: external, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 0, canAutoHide: 0, importType: declaration), insts: 5)))
-; CHECK: ^25 = typeid: (name: "_ZTS1C", summary: (typeTestRes: (kind: single, sizeM1BitWidth: 0))) ; guid = 1884921850105019584
-; CHECK: ^26 = typeid: (name: "_ZTS1B", summary: (typeTestRes: (kind: inline, sizeM1BitWidth: 0, alignLog2: 1, sizeM1: 2, bitMask: 3, inlineBits: 4))) ; guid = 6203814149063363976
-; CHECK: ^27 = typeid: (name: "_ZTS1A", summary: (typeTestRes: (kind: allOnes, sizeM1BitWidth: 7), wpdResolutions: ((offset: 0, wpdRes: (kind: branchFunnel)), (offset: 8, wpdRes: (kind: singleImpl, singleImplName: "_ZN1A1nEi")), (offset: 16, wpdRes: (kind: indir, resByArg: (args: (1, 2), byArg: (kind: indir, byte: 2, bit: 3), args: (3), byArg: (kind: uniformRetVal, info: 1), args: (4), byArg: (kind: uniqueRetVal, info: 1), args: (5), byArg: (kind: virtualConstProp))))))) ; guid = 7004155349499253778
-; CHECK: ^28 = typeid: (name: "_ZTS1D", summary: (typeTestRes: (kind: byteArray, sizeM1BitWidth: 0))) ; guid = 9614786172484273522
-; CHECK: ^29 = typeid: (name: "_ZTS1E", summary: (typeTestRes: (kind: unsat, sizeM1BitWidth: 0))) ; guid = 17437243864166745132
-; CHECK: ^30 = flags: 8
-; CHECK: ^31 = blockcount: 1888
+; CHECK: ^25 = gv: (guid: 9123456789101112131, summaries: (function: (module: ^0, flags: (linkage: internal, visibility: default, notEligibleToImport: 0, live: 0, dsoLocal: 1, canAutoHide: 0, importType: definition), insts: 1)))
+; CHECK: ^26 = typeid: (name: "_ZTS1C", summary: (typeTestRes: (kind: single, sizeM1BitWidth: 0))) ; guid = 1884921850105019584
+; CHECK: ^27 = typeid: (name: "_ZTS1B", summary: (typeTestRes: (kind: inline, sizeM1BitWidth: 0, alignLog2: 1, sizeM1: 2, bitMask: 3, inlineBits: 4))) ; guid = 6203814149063363976
+; CHECK: ^28 = typeid: (name: "_ZTS1A", summary: (typeTestRes: (kind: allOnes, sizeM1BitWidth: 7), wpdResolutions: ((offset: 0, wpdRes: (kind: branchFunnel)), (offset: 8, wpdRes: (kind: singleImpl, singleImplName: "_ZN1A1nEi")), (offset: 16, wpdRes: (kind: indir, resByArg: (args: (1, 2), byArg: (kind: indir, byte: 2, bit: 3), args: (3), byArg: (kind: uniformRetVal, info: 1), args: (4), byArg: (kind: uniqueRetVal, info: 1), args: (5), byArg: (kind: virtualConstProp))))))) ; guid = 7004155349499253778
+; CHECK: ^29 = typeid: (name: "_ZTS1D", summary: (typeTestRes: (kind: byteArray, sizeM1BitWidth: 0))) ; guid = 9614786172484273522
+; CHECK: ^30 = typeid: (name: "_ZTS1E", summary: (typeTestRes: (kind: unsat, sizeM1BitWidth: 0))) ; guid = 17437243864166745132
+; CHECK: ^31 = flags: 8
+; CHECK: ^32 = blockcount: 1888
 
 ; Make sure parsing of a non-summary entry containing a ":" does not fail
 ; after summary parsing, which handles colons differently.
diff --git a/llvm/test/Bitcode/summary_version.ll b/llvm/test/Bitcode/summary_version.ll
index 98feab6fe2f995..26c64f81a773f1 100644
--- a/llvm/test/Bitcode/summary_version.ll
+++ b/llvm/test/Bitcode/summary_version.ll
@@ -2,7 +2,7 @@
 ; RUN: opt  -module-summary  %s -o - | llvm-bcanalyzer -dump | FileCheck %s
 
 ; CHECK: <GLOBALVAL_SUMMARY_BLOCK
-; CHECK: <VERSION op0=9/>
+; CHECK: <VERSION op0=10/>
 
 
 
diff --git a/llvm/test/Bitcode/thinlto-alias.ll b/llvm/test/Bitcode/thinlto-alias.ll
index 5dfff0f796198d..7deb2d8259e328 100644
--- a/llvm/test/Bitcode/thinlto-alias.ll
+++ b/llvm/test/Bitcode/thinlto-alias.ll
@@ -31,9 +31,9 @@
 ; COMBINED-NEXT:    <VERSION
 ; COMBINED-NEXT:    <FLAGS
 ; See if the call to analias is registered, using the expected value id.
-; COMBINED-NEXT:    <VALUE_GUID op0=[[ALIASID:[0-9]+]] op1=-5751648690987223394/>
+; COMBINED-NEXT:    <VALUE_GUID {{.*}} op0=[[ALIASID:[0-9]+]] op1=2955807229 op2=886945438/>
 ; COMBINED-NEXT:    <VALUE_GUID
-; COMBINED-NEXT:    <VALUE_GUID op0=[[ALIASEEID:[0-9]+]] op1=-1039159065113703048/>
+; COMBINED-NEXT:    <VALUE_GUID {{.*}} op0=[[ALIASEEID:[0-9]+]] op1=4053019222 op2=46484856/>
 ; COMBINED-NEXT:    <COMBINED_PROFILE {{.*}} op9=[[ALIASID]]
 ; COMBINED-NEXT:    <COMBINED_PROFILE {{.*}}
 ; COMBINED-NEXT:    <COMBINED_ALIAS  {{.*}} op3=[[ALIASEEID]]
diff --git a/llvm/test/Bitcode/thinlto-func-summary-vtableref-pgo.ll b/llvm/test/Bitcode/thinlto-func-summary-vtableref-pgo.ll
index 19e228fd5355c7..d864cadcff3541 100644
--- a/llvm/test/Bitcode/thinlto-func-summary-vtableref-pgo.ll
+++ b/llvm/test/Bitcode/thinlto-func-summary-vtableref-pgo.ll
@@ -11,21 +11,21 @@
 ; RUN: llvm-dis -o - %t.o | llvm-as -o - | llvm-dis -o - | FileCheck %s --check-prefix=DIS
 
 ; CHECK: <GLOBALVAL_SUMMARY_BLOCK
-; CHECK-NEXT:   <VERSION op0=9/>
+; CHECK-NEXT:   <VERSION op0=10/>
 ; CHECK-NEXT:   <FLAGS op0=0/>
 ; The `VALUE_GUID` below represents the "_ZTV4Base" referenced by the instruction
 ; that loads vtable pointers.
-; CHECK-NEXT: <VALUE_GUID op0=21 op1=1960855528937986108/>
+; CHECK-NEXT: <VALUE_GUID {{.*}} op0=21 op1=456547254 op2=3929380924/>
 ; The `VALUE_GUID` below represents the "_ZN4Base4funcEv" referenced by the
 ; indirect call instruction.
-; CHECK-NEXT:      <VALUE_GUID op0=20 op1=5459407273543877811/>
+; CHECK-NEXT:      <VALUE_GUID {{.*}} op0=20 op1=1271117309 op2=2009351347/>
 ; NOTE vtables and functions from Derived class is dropped because
 ; `-icp-max-num-vtables` and `-icp-max-prom` are both set to one.
 ; <PERMODULE_PROFILE> has the format [valueid, flags, instcount, funcflags,
 ;                                     numrefs, rorefcnt, worefcnt,
 ;                                     m x valueid,
 ;                                     n x (valueid, hotness+tailcall)]
-; CHECK-NEXT:   <PERMODULE_PROFILE abbrevid=4 op0=0 op1=0 op2=4 op3=256 op4=1 op5=1 op6=0 op7=21 op8=20 op9=3/>
+; CHECK-NEXT:   <PERMODULE_PROFILE {{.*}} op0=0 op1=0 op2=4 op3=256 op4=1 op5=1 op6=0 op7=21 op8=20 op9=3/>
 ; CHECK-NEXT:  </GLOBALVAL_SUMMARY_BLOCK>
 
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
diff --git a/llvm/test/Bitcode/thinlto-function-summary-callgraph-partial-sample-profile-summary.ll b/llvm/test/Bitcode/thinlto-function-summary-callgraph-partial-sample-profile-summary.ll
index d44ee24694be56..0c3ab9b2089310 10...
[truncated]

@jvoung
Copy link
Contributor Author

jvoung commented Apr 29, 2024

@teresajohnson

teresajohnson
teresajohnson approved these changes Apr 29, 2024
View reviewed changes
Copy link
Contributor

@teresajohnson teresajohnson left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

thanks! out of curiosity, what happens if you try to specify a 64-bit fixed width value?

@jvoung
Copy link
Contributor Author

jvoung commented Apr 30, 2024

thanks! out of curiosity, what happens if you try to specify a 64-bit fixed width value?

On the writer side, each chunk eventually gets to Emit:

llvm-project/llvm/include/llvm/Bitstream/BitstreamWriter.h

Line 209 in 9a1386e

void Emit(uint32_t Val, unsigned NumBits) {
where the Val is of type uint32_t and there's an assert that requested NumBits is <= 32.

I think there's other assumptions of the accumulator CurValue being 32-bits, and so on. On the Reader side similar checks / assumptions that the MaxChunkSize is 32 bits.

llvm-project/llvm/include/llvm/Bitstream/BitstreamReader.h

Line 378 in 9a1386e

static const size_t MaxChunkSize = 32;

@jvoung
Copy link
Contributor Author

jvoung commented Apr 30, 2024

And thanks for the quick review! Also, I don't have commit access yet/anymore. Would you mind helping merge this PR? Thanks!

@teresajohnson teresajohnson merged commit adabdc1 into llvm:main Apr 30, 2024
8 of 9 checks passed
@teresajohnson
Copy link
Contributor

And thanks for the quick review! Also, I don't have commit access yet/anymore. Would you mind helping merge this PR? Thanks!

Done!

@github-actions GitHub Actions
Copy link

@jvoung Congratulations on having your first Pull Request (PR) merged into the LLVM Project!

Your changes will be combined with recent changes from other authors, then tested
by our build bots. If there is a problem with a build, you may receive a report in an email or a comment on this PR.

Please check whether problems have been caused by your change specifically, as
the builds can include changes from many authors. It is not uncommon for your
change to be included in a build that fails due to someone else's changes, or
infrastructure issues.

How to do this, and the rest of the post-merge process, is covered in detail here.

If your change does cause a problem, it may be reverted, or you can revert it yourself.
This is a normal part of LLVM development. You can fix your changes and open a new PR to merge them again.

If you don't get any reports, no action is required from you. Your changes are working as expected, well done!

@jvoung
Copy link
Contributor Author

jvoung commented Apr 30, 2024

Ah sorry, looks like this broke some LLD MachO and COFF tests that I didn't notice before. I can send a revert then fix.

@jvoung jvoung mentioned this pull request Apr 30, 2024
Merged
@jvoung
Copy link
Contributor Author

jvoung commented Apr 30, 2024

#90610

teresajohnson pushed a commit that referenced this pull request Apr 30, 2024
@jvoung
Revert "Use an abbrev to reduce size of VALUE_GUID records in ThinLTO…
2aabfc8 
… summaries" (#90610)

Reverts #90497
Broke some LLD tests.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@teresajohnson teresajohnson teresajohnson approved these changes

Assignees
No one assigned
Labels
llvm:ir LTO Link time optimization (regular/full LTO or ThinLTO)
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@jvoung @llvmbot @teresajohnson