[Driver] Support fprofile-sample-use= for CL #117282
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
Sampling PGO has already been supported on Windows. This patch adds /fprofile-sample-use= /fprofile-sample-use: /fno-profile-sample-use and supports -fprofile-sample-use= for CL.
llvmbot
added
clang
|
@llvm/pr-subscribers-clang-driver Author: Haohai Wen (HaohaiWen) ChangesSampling PGO has already been supported on Windows. This patch adds Full diff: https://github.com/llvm/llvm-project/pull/117282.diff 3 Files Affected:
diff --git a/clang/docs/UsersManual.rst b/clang/docs/UsersManual.rst
index 2ef603d64557f3..e5d67bc58350c7 100644
--- a/clang/docs/UsersManual.rst
+++ b/clang/docs/UsersManual.rst
@@ -2660,7 +2660,7 @@ usual build cycle when using sample profilers for optimization:
> clang-cl /O2 -gdwarf -gline-tables-only ^
/clang:-fdebug-info-for-profiling /clang:-funique-internal-linkage-names ^
- /fprofile-sample-use=code.prof code.cc /Fe:code /fuse-ld=lld /link /debug:dwarf
+ /fprofile-sample-use=code.prof code.cc /Fe:code -fuse-ld=lld /link /debug:dwarf
[OPTIONAL] Sampling-based profiles can have inaccuracies or missing block/
edge counters. The profile inference algorithm (profi) can be used to infer
@@ -2679,7 +2679,7 @@ usual build cycle when using sample profilers for optimization:
> clang-cl /clang:-fsample-profile-use-profi /O2 -gdwarf -gline-tables-only ^
/clang:-fdebug-info-for-profiling /clang:-funique-internal-linkage-names ^
- /fprofile-sample-use=code.prof code.cc /Fe:code /fuse-ld=lld /link /debug:dwarf
+ /fprofile-sample-use=code.prof code.cc /Fe:code -fuse-ld=lld /link /debug:dwarf
Sample Profile Formats
""""""""""""""""""""""
diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td
index 5167c3c39e315a..3a9655d89ff67b 100644
--- a/clang/include/clang/Driver/Options.td
+++ b/clang/include/clang/Driver/Options.td
@@ -1736,7 +1736,7 @@ def fno_profile_sample_use : Flag<["-"], "fno-profile-sample-use">, Group<f_Grou
Visibility<[ClangOption, CLOption]>;
def fprofile_sample_use_EQ : Joined<["-"], "fprofile-sample-use=">,
Group<f_Group>,
- Visibility<[ClangOption, CC1Option]>,
+ Visibility<[ClangOption, CLOption, CC1Option]>,
HelpText<"Enable sample-based profile guided optimizations">,
MarshallingInfoString<CodeGenOpts<"SampleProfileFile">>;
def fprofile_sample_accurate : Flag<["-"], "fprofile-sample-accurate">,
@@ -8484,6 +8484,9 @@ def _SLASH_fp_strict : CLFlag<"fp:strict">, HelpText<"">, Alias<ffp_model_EQ>, A
def _SLASH_fsanitize_EQ_address : CLFlag<"fsanitize=address">,
HelpText<"Enable AddressSanitizer">,
Alias<fsanitize_EQ>, AliasArgs<["address"]>;
+def : CLJoined<"fno-profile-sample-use">, Alias<fno_profile_sample_use>;
+def : CLJoined<"fprofile-sample-use:">, Alias<fprofile_sample_use_EQ>;
+def : CLJoined<"fprofile-sample-use=">, Alias<fprofile_sample_use_EQ>;
def _SLASH_GA : CLFlag<"GA">, Alias<ftlsmodel_EQ>, AliasArgs<["local-exec"]>,
HelpText<"Assume thread-local variables are defined in the executable">;
def _SLASH_GR : CLFlag<"GR">, HelpText<"Emit RTTI data (default)">;
diff --git a/clang/test/Driver/cl-options.c b/clang/test/Driver/cl-options.c
index 8191fda97788c1..f219304111a94c 100644
--- a/clang/test/Driver/cl-options.c
+++ b/clang/test/Driver/cl-options.c
@@ -101,6 +101,14 @@
// CHECK-PROFILE-USE: "-fprofile-instrument-use-path=default.profdata"
// CHECK-PROFILE-USE-FILE: "-fprofile-instrument-use-path=/tmp/somefile.prof"
+// RUN: %clang_cl -### /FA -fprofile-sample-use=%S/Inputs/file.prof -- %s 2>&1 | FileCheck --check-prefix=CHECK-PROFILE-SAMPLE-USE %s
+// RUN: %clang_cl -### /FA /fprofile-sample-use=%S/Inputs/file.prof -- %s 2>&1 | FileCheck --check-prefix=CHECK-PROFILE-SAMPLE-USE %s
+// RUN: %clang_cl -### /FA /fprofile-sample-use:%S/Inputs/file.prof -- %s 2>&1 | FileCheck --check-prefix=CHECK-PROFILE-SAMPLE-USE %s
+// CHECK-PROFILE-SAMPLE-USE: "-fprofile-sample-use={{.*}}/file.prof"
+
+// RUN: %clang_cl -### /FA /fprofile-sample-use=%S/Inputs/file.prof /fno-profile-sample-use -- %s 2>&1 | FileCheck -check-prefix=CHECK-NO-PROFILE-SAMPLE-USE %s
+// CHECK-NO-PROFILE-SAMPLE-USE-NOT: "-fprofile-sample-use={{.*}}/file.prof"
+
// RUN: %clang_cl /GA -### -- %s 2>&1 | FileCheck -check-prefix=GA %s
// GA: -ftls-model=local-exec
|
|
@llvm/pr-subscribers-clang Author: Haohai Wen (HaohaiWen) ChangesSampling PGO has already been supported on Windows. This patch adds Full diff: https://github.com/llvm/llvm-project/pull/117282.diff 3 Files Affected:
diff --git a/clang/docs/UsersManual.rst b/clang/docs/UsersManual.rst
index 2ef603d64557f3..e5d67bc58350c7 100644
--- a/clang/docs/UsersManual.rst
+++ b/clang/docs/UsersManual.rst
@@ -2660,7 +2660,7 @@ usual build cycle when using sample profilers for optimization:
> clang-cl /O2 -gdwarf -gline-tables-only ^
/clang:-fdebug-info-for-profiling /clang:-funique-internal-linkage-names ^
- /fprofile-sample-use=code.prof code.cc /Fe:code /fuse-ld=lld /link /debug:dwarf
+ /fprofile-sample-use=code.prof code.cc /Fe:code -fuse-ld=lld /link /debug:dwarf
[OPTIONAL] Sampling-based profiles can have inaccuracies or missing block/
edge counters. The profile inference algorithm (profi) can be used to infer
@@ -2679,7 +2679,7 @@ usual build cycle when using sample profilers for optimization:
> clang-cl /clang:-fsample-profile-use-profi /O2 -gdwarf -gline-tables-only ^
/clang:-fdebug-info-for-profiling /clang:-funique-internal-linkage-names ^
- /fprofile-sample-use=code.prof code.cc /Fe:code /fuse-ld=lld /link /debug:dwarf
+ /fprofile-sample-use=code.prof code.cc /Fe:code -fuse-ld=lld /link /debug:dwarf
Sample Profile Formats
""""""""""""""""""""""
diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td
index 5167c3c39e315a..3a9655d89ff67b 100644
--- a/clang/include/clang/Driver/Options.td
+++ b/clang/include/clang/Driver/Options.td
@@ -1736,7 +1736,7 @@ def fno_profile_sample_use : Flag<["-"], "fno-profile-sample-use">, Group<f_Grou
Visibility<[ClangOption, CLOption]>;
def fprofile_sample_use_EQ : Joined<["-"], "fprofile-sample-use=">,
Group<f_Group>,
- Visibility<[ClangOption, CC1Option]>,
+ Visibility<[ClangOption, CLOption, CC1Option]>,
HelpText<"Enable sample-based profile guided optimizations">,
MarshallingInfoString<CodeGenOpts<"SampleProfileFile">>;
def fprofile_sample_accurate : Flag<["-"], "fprofile-sample-accurate">,
@@ -8484,6 +8484,9 @@ def _SLASH_fp_strict : CLFlag<"fp:strict">, HelpText<"">, Alias<ffp_model_EQ>, A
def _SLASH_fsanitize_EQ_address : CLFlag<"fsanitize=address">,
HelpText<"Enable AddressSanitizer">,
Alias<fsanitize_EQ>, AliasArgs<["address"]>;
+def : CLJoined<"fno-profile-sample-use">, Alias<fno_profile_sample_use>;
+def : CLJoined<"fprofile-sample-use:">, Alias<fprofile_sample_use_EQ>;
+def : CLJoined<"fprofile-sample-use=">, Alias<fprofile_sample_use_EQ>;
def _SLASH_GA : CLFlag<"GA">, Alias<ftlsmodel_EQ>, AliasArgs<["local-exec"]>,
HelpText<"Assume thread-local variables are defined in the executable">;
def _SLASH_GR : CLFlag<"GR">, HelpText<"Emit RTTI data (default)">;
diff --git a/clang/test/Driver/cl-options.c b/clang/test/Driver/cl-options.c
index 8191fda97788c1..f219304111a94c 100644
--- a/clang/test/Driver/cl-options.c
+++ b/clang/test/Driver/cl-options.c
@@ -101,6 +101,14 @@
// CHECK-PROFILE-USE: "-fprofile-instrument-use-path=default.profdata"
// CHECK-PROFILE-USE-FILE: "-fprofile-instrument-use-path=/tmp/somefile.prof"
+// RUN: %clang_cl -### /FA -fprofile-sample-use=%S/Inputs/file.prof -- %s 2>&1 | FileCheck --check-prefix=CHECK-PROFILE-SAMPLE-USE %s
+// RUN: %clang_cl -### /FA /fprofile-sample-use=%S/Inputs/file.prof -- %s 2>&1 | FileCheck --check-prefix=CHECK-PROFILE-SAMPLE-USE %s
+// RUN: %clang_cl -### /FA /fprofile-sample-use:%S/Inputs/file.prof -- %s 2>&1 | FileCheck --check-prefix=CHECK-PROFILE-SAMPLE-USE %s
+// CHECK-PROFILE-SAMPLE-USE: "-fprofile-sample-use={{.*}}/file.prof"
+
+// RUN: %clang_cl -### /FA /fprofile-sample-use=%S/Inputs/file.prof /fno-profile-sample-use -- %s 2>&1 | FileCheck -check-prefix=CHECK-NO-PROFILE-SAMPLE-USE %s
+// CHECK-NO-PROFILE-SAMPLE-USE-NOT: "-fprofile-sample-use={{.*}}/file.prof"
+
// RUN: %clang_cl /GA -### -- %s 2>&1 | FileCheck -check-prefix=GA %s
// GA: -ftls-model=local-exec
|
| @@ -2660,7 +2660,7 @@ usual build cycle when using sample profilers for optimization: | |||
|
|
|||
| > clang-cl /O2 -gdwarf -gline-tables-only ^ | |||
| /clang:-fdebug-info-for-profiling /clang:-funique-internal-linkage-names ^ | |||
| /fprofile-sample-use=code.prof code.cc /Fe:code /fuse-ld=lld /link /debug:dwarf | |||
There was a problem hiding this comment.
I take it this wasn't working? (Neither /fprofile-sample-use nor /fuse-ld?)
There was a problem hiding this comment.
Yes. I checked options.td.
-fprofile-sample-use was not a CLOption and there's no /fuse-ld.
HaohaiWen
added
the
clang-cl
MaskRay
reviewed
Nov 25, 2024
| @@ -8484,6 +8484,9 @@ def _SLASH_fp_strict : CLFlag<"fp:strict">, HelpText<"">, Alias<ffp_model_EQ>, A | |||
| def _SLASH_fsanitize_EQ_address : CLFlag<"fsanitize=address">, | |||
| HelpText<"Enable AddressSanitizer">, | |||
| Alias<fsanitize_EQ>, AliasArgs<["address"]>; | |||
| def : CLJoined<"fno-profile-sample-use">, Alias<fno_profile_sample_use>; | |||
There was a problem hiding this comment.
Do we really need these aliases?
There was a problem hiding this comment.
For me, /fprofile-sample-use= is needed.
I saw other CL _EQ flags has _COL alias.
In fact, I don't really need /fno-profile-sample-use.
There was a problem hiding this comment.
Removed /fno-profile-sample-use.
|
I'd like to merge it in if no objection. |
|
Adding CLOption to -fprofile-sample-use= suffices. |
Got it. I checked CL flags with MSVC. You're right. |
HaohaiWen
added a commit
to HaohaiWen/llvm-project
that referenced
this pull request
Nov 28, 2024
Those flags are introduced in llvm#117282. They are not supported by MSVC.
HaohaiWen
added a commit
that referenced
this pull request
Nov 28, 2024
Those flags are introduced in #117282. They are not supported by MSVC.
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.
Sampling PGO has already been supported on Windows. This patch adds
/fprofile-sample-use= /fprofile-sample-use: /fno-profile-sample-use and
supports -fprofile-sample-use= for CL.