Pass -foffload-lto instead of -flto for cuda/hip kernels in clangLinkerWrapper #16605
Conversation
This file contains 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
omarahmed1111
force-pushed
the
fix-NewOffloadDrivers-failed-tests
branch
from
5b77ca9 to
d24ae0e
Compare
omarahmed1111
had a problem deploying
to
WindowsCILock
GitHub Actions
Error
omarahmed1111
force-pushed
the
fix-NewOffloadDrivers-failed-tests
branch
from
d24ae0e to
8615f88
Compare
omarahmed1111
temporarily deployed
to
WindowsCILock
GitHub Actions
Inactive
omarahmed1111
had a problem deploying
to
WindowsCILock
GitHub Actions
Error
omarahmed1111
force-pushed
the
fix-NewOffloadDrivers-failed-tests
branch
from
8615f88 to
6c7fbf7
Compare
omarahmed1111
had a problem deploying
to
WindowsCILock
GitHub Actions
Error
omarahmed1111
force-pushed
the
fix-NewOffloadDrivers-failed-tests
branch
from
6c7fbf7 to
da775da
Compare
omarahmed1111
had a problem deploying
to
WindowsCILock
GitHub Actions
Error
omarahmed1111
force-pushed
the
fix-NewOffloadDrivers-failed-tests
branch
from
da775da to
633d90b
Compare
omarahmed1111
had a problem deploying
to
WindowsCILock
GitHub Actions
Error
omarahmed1111
force-pushed
the
fix-NewOffloadDrivers-failed-tests
branch
from
633d90b to
ccf221a
Compare
omarahmed1111
temporarily deployed
to
WindowsCILock
GitHub Actions
Inactive
omarahmed1111
temporarily deployed
to
WindowsCILock
GitHub Actions
Inactive
sarnex
reviewed
Jan 14, 2025
| for (auto &Arg : Args.filtered(OPT_offload_opt_eq_minus, OPT_mllvm)) | ||
| CmdArgs.append( | ||
| {"-Xlinker", | ||
| Args.MakeArgString("--plugin-opt=" + StringRef(Arg->getValue()))}); | ||
|
|
||
| if (Triple.isNVPTX() || Triple.isAMDGPU()) { | ||
| CmdArgs.push_back("-foffload-lto"); |
There was a problem hiding this comment.
+1 on adding this change to upstream code. I am also wondering why it is not an issue in the upstream compiler and why we see it only with SYCL offloading.
Also @omarahmed1111 can you please comment on the test failure?
Thanks
There was a problem hiding this comment.
The test failure for loop_extended is a flaky failure not related to this PR, the failed test disabled on this PR: #16612 . I updated the PR, so should be solved now.
omarahmed1111
temporarily deployed
to
WindowsCILock
GitHub Actions
Inactive
omarahmed1111
temporarily deployed
to
WindowsCILock
GitHub Actions
Inactive
sarnex
reviewed
Jan 27, 2025
| for (auto &Arg : Args.filtered(OPT_offload_opt_eq_minus, OPT_mllvm)) | ||
| CmdArgs.append( | ||
| {"-Xlinker", | ||
| Args.MakeArgString("--plugin-opt=" + StringRef(Arg->getValue()))}); | ||
|
|
||
| if (Triple.isNVPTX() || Triple.isAMDGPU()) { |
There was a problem hiding this comment.
is there an upstream pr for this based on previous feedback?
There was a problem hiding this comment.
I thought we are going to do this after merging this. I will open one for that.
There was a problem hiding this comment.
Ok as long as you'll make an upstream PR lgtm to merge here first
sarnex
approved these changes
Jan 28, 2025
mdtoguchi
approved these changes
Jan 28, 2025
|
@intel/llvm-gatekeepers please merge, Thanks! |
|
@sarnex llvm upstream PR: llvm/llvm-project#125243 |
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.
ClangLinkerWrapper tool in one of its clang commands to generate ptx kernel binary from llvm bitcode kernel was using
-fltooption which should be only used for cpu code not gpu kernel code. This PR fixes that by changing that to-foffload-ltofor cuda/hip kernels.This fixes 16413 issue.