[CMake][PGO] Add option for using an external project to generate profile data #78879

tstellar · 2024-01-21T05:51:27Z

The new CLANG_PGO_TRAINING_DATA_SOURCE_DIR allows users to specify a CMake project to use for generating the profile data. For example, to use the llvm-test-suite to generate profile data you would do:

$ cmake -G Ninja -B build -S llvm -C /clang/cmake/caches/PGO.cmake \
-DBOOTSTRAP_CLANG_PGO_TRAINING_DATA_SOURCE_DIR=
-DBOOTSTRAP_CLANG_PGO_TRAINING_DEPS=runtimes

Note that the CLANG_PERF_TRAINING_DEPS has been renamed to CLANG_PGO_TRAINING_DEPS.

This fixes the build since 8f90e69 which made libcxxabi use llvm's libunwind by default.

…file data. The new CLANG_PGO_TRAINING_DATA_SOURCE_DIR allows users to specify a CMake project to use for generating the profile data. For example, to use the llvm-test-suite to generate profile data you would do: $ cmake -G Ninja -B build -S llvm -C <path to source>/clang/cmake/caches/PGO.cmake \ -DBOOTSTRAP_CLANG_PGO_TRAINING_DATA_SOURCE_DIR=<path to llvm-test-suite> \ -DBOOTSTRAP_CLANG_PERF_TRAINING_DEPS=runtimes

llvmbot · 2024-01-21T05:52:02Z

@llvm/pr-subscribers-clang

Author: Tom Stellard (tstellar)

Changes

The new CLANG_PGO_TRAINING_DATA_SOURCE_DIR allows users to specify a CMake project to use for generating the profile data. For example, to use the llvm-test-suite to generate profile data you would do:

$ cmake -G Ninja -B build -S llvm -C <path to source>/clang/cmake/caches/PGO.cmake \
-DBOOTSTRAP_CLANG_PGO_TRAINING_DATA_SOURCE_DIR=<path to llvm-test-suite>
-DBOOTSTRAP_CLANG_PERF_TRAINING_DEPS=runtimes

Full diff: https://github.com/llvm/llvm-project/pull/78879.diff

4 Files Affected:

(modified) clang/cmake/caches/PGO.cmake (+1-1)
(modified) clang/utils/perf-training/CMakeLists.txt (+11-2)
(modified) clang/utils/perf-training/perf-helper.py (+9-7)
(modified) llvm/docs/AdvancedBuilds.rst (+23)

diff --git a/clang/cmake/caches/PGO.cmake b/clang/cmake/caches/PGO.cmake
index e1d0585e453f825..15bc755d110d19a 100644
--- a/clang/cmake/caches/PGO.cmake
+++ b/clang/cmake/caches/PGO.cmake
@@ -2,7 +2,7 @@ set(CMAKE_BUILD_TYPE RELEASE CACHE STRING "")
 set(CLANG_ENABLE_BOOTSTRAP ON CACHE BOOL "")
 
 set(LLVM_ENABLE_PROJECTS "clang;lld" CACHE STRING "")
-set(LLVM_ENABLE_RUNTIMES "compiler-rt;libcxx;libcxxabi" 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 "")
diff --git a/clang/utils/perf-training/CMakeLists.txt b/clang/utils/perf-training/CMakeLists.txt
index c6d51863fb1b5c2..f9d673b2e92e775 100644
--- a/clang/utils/perf-training/CMakeLists.txt
+++ b/clang/utils/perf-training/CMakeLists.txt
@@ -1,6 +1,10 @@
+include(LLVMExternalProjectUtils)
+
 set(CLANG_PGO_TRAINING_DATA "${CMAKE_CURRENT_SOURCE_DIR}" CACHE PATH
   "The path to a lit testsuite containing samples for PGO and order file generation"
   )
+set(CLANG_PGO_TRAINING_DATA_SOURCE_DIR OFF CACHE STRING "Path to source directory containing cmake project with source files to use for generating pgo data")
+set(CLANG_PERF_TRAINING_DEPS "" CACHE STRING "Extra dependencies needed to build the PGO training data.")
 
 if(LLVM_BUILD_INSTRUMENTED)
   configure_lit_site_cfg(
@@ -15,7 +19,7 @@ if(LLVM_BUILD_INSTRUMENTED)
     )
 
   add_custom_target(clear-profraw
-    COMMAND "${Python3_EXECUTABLE}" ${CMAKE_CURRENT_SOURCE_DIR}/perf-helper.py clean ${CMAKE_CURRENT_BINARY_DIR} profraw
+    COMMAND "${Python3_EXECUTABLE}" ${CMAKE_CURRENT_SOURCE_DIR}/perf-helper.py clean ${CMAKE_CURRENT_BINARY_DIR} ${CMAKE_BINARY_DIR}/profiles/ profraw
     COMMENT "Clearing old profraw data")
 
   if(NOT LLVM_PROFDATA)
@@ -26,9 +30,14 @@ if(LLVM_BUILD_INSTRUMENTED)
     message(STATUS "To enable merging PGO data LLVM_PROFDATA has to point to llvm-profdata")
   else()
     add_custom_target(generate-profdata
-      COMMAND "${Python3_EXECUTABLE}" ${CMAKE_CURRENT_SOURCE_DIR}/perf-helper.py merge ${LLVM_PROFDATA} ${CMAKE_CURRENT_BINARY_DIR}/clang.profdata ${CMAKE_CURRENT_BINARY_DIR}
+      COMMAND "${Python3_EXECUTABLE}" ${CMAKE_CURRENT_SOURCE_DIR}/perf-helper.py merge ${LLVM_PROFDATA} ${CMAKE_CURRENT_BINARY_DIR}/clang.profdata ${CMAKE_CURRENT_BINARY_DIR} ${CMAKE_BINARY_DIR}/profiles/
       COMMENT "Merging profdata"
       DEPENDS generate-profraw)
+    if (CLANG_PGO_TRAINING_DATA_SOURCE_DIR)
+      llvm_ExternalProject_Add(generate-profraw-external ${CLANG_PGO_TRAINING_DATA_SOURCE_DIR}
+              USE_TOOLCHAIN EXLUDE_FROM_ALL NO_INSTALL DEPENDS generate-profraw)
+      add_dependencies(generate-profdata generate-profraw-external)
+    endif()
   endif()
 endif()
 
diff --git a/clang/utils/perf-training/perf-helper.py b/clang/utils/perf-training/perf-helper.py
index 99d6a3333b6ef08..844aa274f049aaa 100644
--- a/clang/utils/perf-training/perf-helper.py
+++ b/clang/utils/perf-training/perf-helper.py
@@ -30,26 +30,28 @@ def findFilesWithExtension(path, extension):
 
 
 def clean(args):
-    if len(args) != 2:
+    if len(args) < 2:
         print(
-            "Usage: %s clean <path> <extension>\n" % __file__
+            "Usage: %s clean <paths> <extension>\n" % __file__
             + "\tRemoves all files with extension from <path>."
         )
         return 1
-    for filename in findFilesWithExtension(args[0], args[1]):
-        os.remove(filename)
+    for path in args[1:-1]:
+        for filename in findFilesWithExtension(path, args[-1]):
+            os.remove(filename)
     return 0
 
 
 def merge(args):
-    if len(args) != 3:
+    if len(args) < 3:
         print(
-            "Usage: %s merge <llvm-profdata> <output> <path>\n" % __file__
+            "Usage: %s merge <llvm-profdata> <output> <paths>\n" % __file__
             + "\tMerges all profraw files from path into output."
         )
         return 1
     cmd = [args[0], "merge", "-o", args[1]]
-    cmd.extend(findFilesWithExtension(args[2], "profraw"))
+    for i in range(2, len(args)):
+        cmd.extend(findFilesWithExtension(args[i], "profraw"))
     subprocess.check_call(cmd)
     return 0
 
diff --git a/llvm/docs/AdvancedBuilds.rst b/llvm/docs/AdvancedBuilds.rst
index 960b19fa5317f3b..3ff630a1425691d 100644
--- a/llvm/docs/AdvancedBuilds.rst
+++ b/llvm/docs/AdvancedBuilds.rst
@@ -145,6 +145,29 @@ that also enables ThinTLO, use the following command:
       -DPGO_INSTRUMENT_LTO=Thin \
       <path to source>/llvm
 
+By default, clang will generate profile data by compiling a simple
+hello world program.  You can also tell clang use an external
+project for generating profile data that may be a better fit for your
+use case.  The project you specify must either be a lit test suite
+(use the CLANG_PGO_TRAINING_DATA option) or a CMake project (use the
+CLANG_PERF_TRAINING_DATA_SOURCE_DIR option).
+
+For example, If you wanted to use the
+`LLVM Test Suite <https://github.com/llvm/llvm-test-suite/>`_ to generate
+profile data you would use the following command:
+
+.. code-block:: console
+
+  $ cmake -G Ninja -C <path to source>/clang/cmake/caches/PGO.cmake \
+       -DBOOTSTRAP_CLANG_PGO_TRAINING_DATA_SOURCE_DIR=<path to llvm-test-suite> \
+       -DBOOTSTRAP_CLANG_PERF_TRAINING_DEPS=runtimes
+
+The BOOTSTRAP\_ prefixes tells CMake to pass the variables on to the instrumented
+stage two build.  And the CLANG_PERF_TRAINING_DEPS option let's you specify
+additional build targets to build before building the external project.  The
+LLVM Test Suite requires compiler-rt to build, so we need to add the
+`runtimes` target as a dependency.
+
 After configuration, building the stage2-instrumented-generate-profdata target
 will automatically build the stage1 compiler, build the instrumented compiler
 with the stage1 compiler, and then run the instrumented compiler against the

kwk

LGTM

llvm-beanz

LGTM

petrhosek · 2024-01-23T16:48:10Z

clang/utils/perf-training/CMakeLists.txt

 set(CLANG_PGO_TRAINING_DATA "${CMAKE_CURRENT_SOURCE_DIR}" CACHE PATH
  "The path to a lit testsuite containing samples for PGO and order file generation"
  )
+set(CLANG_PGO_TRAINING_DATA_SOURCE_DIR OFF CACHE STRING "Path to source directory containing cmake project with source files to use for generating pgo data")
+set(CLANG_PERF_TRAINING_DEPS "" CACHE STRING "Extra dependencies needed to build the PGO training data.")


I don't see this variable being used anywhere, should it be added as a dependency of generate-profraw-external? I'd also consider naming it CLANG_PGO_TRAINING_DEPS for consistency with other variables in this file.

clang/utils/perf-training/perf-helper.py

tstellar · 2024-01-23T19:26:46Z

clang/utils/perf-training/CMakeLists.txt

@@ -15,7 +19,7 @@ if(LLVM_BUILD_INSTRUMENTED)
    )


@petrhosek CLANG_PERF_TRAINING_DEPS was an existing variable that's used here. I think it was meant to be added to cache files, but I added it to CMakeLists.txt file here so that it could be passed on the command line without modifying the cache files.

Looks like I added that in https://reviews.llvm.org/D138974 but I'm not aware of it being used anywhere at the moment. I'd still prefer renaming it to CLANG_PGO_TRAINING_DEPS for consistency, especially that we're now adding it to cache.

Co-authored-by: Petr Hosek <phosek@google.com>

aaupov · 2024-01-30T21:05:47Z

I'm in favor of this change especially in context of CSSPGO profiling (#79942) and BOLT LBR profiling (#69133).

aaupov · 2024-01-30T21:12:19Z

clang/cmake/caches/PGO.cmake

@@ -2,7 +2,7 @@ set(CMAKE_BUILD_TYPE RELEASE CACHE STRING "")
 set(CLANG_ENABLE_BOOTSTRAP ON CACHE BOOL "")

 set(LLVM_ENABLE_PROJECTS "clang;lld" CACHE STRING "")
-set(LLVM_ENABLE_RUNTIMES "compiler-rt;libcxx;libcxxabi" CACHE STRING "")
+set(LLVM_ENABLE_RUNTIMES "compiler-rt;libcxx;libcxxabi;libunwind" CACHE STRING "")


Makes sense to avoid touching to prevent conflicts with #78869.

@aaupov How does this look now?

aaupov · 2024-01-30T21:17:30Z

clang/utils/perf-training/CMakeLists.txt

    )

  add_custom_target(clear-profraw
-    COMMAND "${Python3_EXECUTABLE}" ${CMAKE_CURRENT_SOURCE_DIR}/perf-helper.py clean ${CMAKE_CURRENT_BINARY_DIR} profraw
+    COMMAND "${Python3_EXECUTABLE}" ${CMAKE_CURRENT_SOURCE_DIR}/perf-helper.py clean ${CMAKE_CURRENT_BINARY_DIR} ${CMAKE_BINARY_DIR}/profiles/ profraw


Where does this new path (${CMAKE_BINARY_DIR}/profiles/) come from?

@aaupov That's the directory where clang will write it's profiles to by default.

aaupov

LGTM

…file data (llvm#78879) The new CLANG_PGO_TRAINING_DATA_SOURCE_DIR allows users to specify a CMake project to use for generating the profile data. For example, to use the llvm-test-suite to generate profile data you would do: $ cmake -G Ninja -B build -S llvm -C <path to source>/clang/cmake/caches/PGO.cmake \ -DBOOTSTRAP_CLANG_PGO_TRAINING_DATA_SOURCE_DIR=<path to llvm-test-suite> \ -DBOOTSTRAP_CLANG_PGO_TRAINING_DEPS=runtimes Note that the CLANG_PERF_TRAINING_DEPS has been renamed to CLANG_PGO_TRAINING_DEPS. --------- Co-authored-by: Petr Hosek <phosek@google.com> (cherry picked from commit dd0356d)

…file data (llvm#78879) The new CLANG_PGO_TRAINING_DATA_SOURCE_DIR allows users to specify a CMake project to use for generating the profile data. For example, to use the llvm-test-suite to generate profile data you would do: $ cmake -G Ninja -B build -S llvm -C <path to source>/clang/cmake/caches/PGO.cmake \ -DBOOTSTRAP_CLANG_PGO_TRAINING_DATA_SOURCE_DIR=<path to llvm-test-suite> \ -DBOOTSTRAP_CLANG_PGO_TRAINING_DEPS=runtimes Note that the CLANG_PERF_TRAINING_DEPS has been renamed to CLANG_PGO_TRAINING_DEPS. --------- Co-authored-by: Petr Hosek <phosek@google.com>

…file data (llvm#78879) The new CLANG_PGO_TRAINING_DATA_SOURCE_DIR allows users to specify a CMake project to use for generating the profile data. For example, to use the llvm-test-suite to generate profile data you would do: $ cmake -G Ninja -B build -S llvm -C <path to source>/clang/cmake/caches/PGO.cmake \ -DBOOTSTRAP_CLANG_PGO_TRAINING_DATA_SOURCE_DIR=<path to llvm-test-suite> \ -DBOOTSTRAP_CLANG_PGO_TRAINING_DEPS=runtimes Note that the CLANG_PERF_TRAINING_DEPS has been renamed to CLANG_PGO_TRAINING_DEPS. --------- Co-authored-by: Petr Hosek <phosek@google.com> (cherry picked from commit dd0356d)

tstellar added 2 commits January 21, 2024 01:26

[CMake][PGO] Add libunwind to list of stage1 runtimes

0719f49

This fixes the build since 8f90e69 which made libcxxabi use llvm's libunwind by default.

tstellar requested review from petrhosek and llvm-beanz January 21, 2024 05:51

llvmbot added the clang Clang issues not falling into any other category label Jan 21, 2024

tstellar mentioned this pull request Jan 21, 2024

[CMake][PGO] Build Sema.cpp to generate profdata for PGO builds #77347

Open

kwk approved these changes Jan 23, 2024

View reviewed changes

llvm-beanz approved these changes Jan 23, 2024

View reviewed changes

petrhosek reviewed Jan 23, 2024

View reviewed changes

clang/utils/perf-training/perf-helper.py Outdated Show resolved Hide resolved

tstellar commented Jan 23, 2024

View reviewed changes

tstellar and others added 2 commits January 23, 2024 11:29

Update clang/utils/perf-training/perf-helper.py

4f9fa29

Co-authored-by: Petr Hosek <phosek@google.com>

Rename CLANG_PERF_TRAINING_DEPS to CLANG_PGO_TRAINING_DEPS

552d251

petrhosek approved these changes Jan 24, 2024

View reviewed changes

boomanaiden154 mentioned this pull request Jan 30, 2024

[Clang][CMake] Add CSSPGO support to LLVM_BUILD_INSTRUMENTED #79942

Open

aaupov reviewed Jan 30, 2024

View reviewed changes

tstellar mentioned this pull request Jan 31, 2024

[Github] Build PGO optimized toolchain in container #80096

Merged

Merge remote-tracking branch 'origin/main' into clang-pgo-external

80d98a7

aaupov approved these changes Feb 2, 2024

View reviewed changes

tstellar merged commit dd0356d into llvm:main Feb 2, 2024
4 of 5 checks passed

pointhex mentioned this pull request May 7, 2024

getStyleDiagHandler #91314

Closed

aemerson mentioned this pull request May 9, 2024

release/18.x: [AArc64][GlobalISel] Fix legalizer assert for G_INSERT_VECTOR_ELT - manual merge #91672

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[CMake][PGO] Add option for using an external project to generate profile data #78879

[CMake][PGO] Add option for using an external project to generate profile data #78879

tstellar commented Jan 21, 2024 •

edited

Loading

llvmbot commented Jan 21, 2024

kwk left a comment

llvm-beanz left a comment

petrhosek Jan 23, 2024

tstellar Jan 23, 2024

petrhosek Jan 23, 2024

aaupov commented Jan 30, 2024

aaupov Jan 30, 2024

tstellar Feb 2, 2024

aaupov Jan 30, 2024

tstellar Jan 30, 2024

aaupov left a comment

[CMake][PGO] Add option for using an external project to generate profile data #78879

[CMake][PGO] Add option for using an external project to generate profile data #78879

Conversation

tstellar commented Jan 21, 2024 • edited Loading

llvmbot commented Jan 21, 2024

kwk left a comment

Choose a reason for hiding this comment

llvm-beanz left a comment

Choose a reason for hiding this comment

petrhosek Jan 23, 2024

Choose a reason for hiding this comment

tstellar Jan 23, 2024

Choose a reason for hiding this comment

petrhosek Jan 23, 2024

Choose a reason for hiding this comment

aaupov commented Jan 30, 2024

aaupov Jan 30, 2024

Choose a reason for hiding this comment

tstellar Feb 2, 2024

Choose a reason for hiding this comment

aaupov Jan 30, 2024

Choose a reason for hiding this comment

tstellar Jan 30, 2024

Choose a reason for hiding this comment

aaupov left a comment

Choose a reason for hiding this comment

tstellar commented Jan 21, 2024 •

edited

Loading