[Clang][CMake] Use IRPGO instead of FE PGO for Cmake Caches #155957
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
Currently the clang CMake caches use FE PGO for instrumentation (LLVM_BUILD_INSTRUMENTED=ON). However, IRPGO is generally regarded as better for performance. I am measuring about a 1.5% performance gain when building libLLVMSupport.a using this configuration versus what existed before this commit. I would suspect the gains are larger on other platforms like Windows where we cannot subsume any gains using PLO.
|
@llvm/pr-subscribers-clang Author: Aiden Grossman (boomanaiden154) ChangesCurrently the clang CMake caches use FE PGO for instrumentation (LLVM_BUILD_INSTRUMENTED=ON). However, IRPGO is generally regarded as better for performance. I am measuring about a 1.5% performance gain when building libLLVMSupport.a using this configuration versus what existed before this commit. I would suspect the gains are larger on other platforms like Windows where we cannot subsume any gains using PLO. Full diff: https://github.com/llvm/llvm-project/pull/155957.diff 1 Files Affected:
diff --git a/clang/cmake/caches/PGO.cmake b/clang/cmake/caches/PGO.cmake
index 15bc755d110d1..d6471160037c1 100644
--- a/clang/cmake/caches/PGO.cmake
+++ b/clang/cmake/caches/PGO.cmake
@@ -5,7 +5,7 @@ set(LLVM_ENABLE_PROJECTS "clang;lld" CACHE STRING "")
set(LLVM_ENABLE_RUNTIMES "compiler-rt;libcxx;libcxxabi;libunwind" CACHE STRING "")
set(LLVM_TARGETS_TO_BUILD Native CACHE STRING "")
-set(BOOTSTRAP_LLVM_BUILD_INSTRUMENTED ON CACHE BOOL "")
+set(BOOTSTRAP_LLVM_BUILD_INSTRUMENTED IR CACHE BOOL "")
set(CLANG_BOOTSTRAP_TARGETS
generate-profdata
stage2
|
mtrofin
approved these changes
Aug 29, 2025
ilovepi
added a commit
that referenced
this pull request
Aug 29, 2025
After #155957 pointed it out, we're actually using the frontend PGO in the Fuchsia cmake cache files.
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/10/builds/12477 Here is the relevant piece of the build log for the reference |
ilovepi
added a commit
that referenced
this pull request
Aug 29, 2025
After #155957 pointed it out, we're actually using the frontend PGO in the Fuchsia cmake cache files.
|
/cherry-pick 7fca1f8 |
|
/pull-request #156271 |
tru
pushed a commit
to llvmbot/llvm-project
that referenced
this pull request
Sep 3, 2025
) Currently the clang CMake caches use FE PGO for instrumentation (LLVM_BUILD_INSTRUMENTED=ON). However, IRPGO is generally regarded as better for performance. I am measuring about a 1.5% performance gain when building libLLVMSupport.a using this configuration versus what existed before this commit. I would suspect the gains are larger on other platforms like Windows where we cannot subsume any gains using PLO. (cherry picked from commit 7fca1f8)
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.
Currently the clang CMake caches use FE PGO for instrumentation (LLVM_BUILD_INSTRUMENTED=ON). However, IRPGO is generally regarded as better for performance. I am measuring about a 1.5% performance gain when building libLLVMSupport.a using this configuration versus what existed before this commit. I would suspect the gains are larger on other platforms like Windows where we cannot subsume any gains using PLO.