Skip to content
This repository was archived by the owner on Mar 28, 2023. It is now read-only.

[SYCL] Add support for the generic address space in tests for sycl::a… #619

Merged
merged 8 commits into from Dec 23, 2021
Merged

[SYCL] Add support for the generic address space in tests for sycl::a… #619

merged 8 commits into from Dec 23, 2021

Conversation

ghost
Copy link

@ghost ghost commented Dec 15, 2021

…tomic_ref

Signed-off-by: Pavel Samolysov pavel.samolysov@intel.com

@ghost
Copy link
Author

ghost commented Dec 15, 2021

/verify with intel/llvm#5148

5 similar comments
@ghost
Copy link
Author

ghost commented Dec 16, 2021

/verify with intel/llvm#5148

@ghost
Copy link
Author

ghost commented Dec 16, 2021

/verify with intel/llvm#5148

@ghost
Copy link
Author

ghost commented Dec 17, 2021

/verify with intel/llvm#5148

@ghost
Copy link
Author

ghost commented Dec 17, 2021

/verify with intel/llvm#5148

@ghost
Copy link
Author

ghost commented Dec 17, 2021

/verify with intel/llvm#5148

@ghost
Copy link
Author

ghost commented Dec 17, 2021

Very strange, during the previous run, CUDA job was not able to get the changes from git.

@ghost
Copy link
Author

ghost commented Dec 20, 2021

/verify with intel/llvm#5148

@ghost
Copy link
Author

ghost commented Dec 21, 2021

Some tests sporadically failed. Re-run the tests against the feature commit in sycl.

@ghost
Copy link
Author

ghost commented Dec 21, 2021

/verify with intel/llvm#5148

1 similar comment
@ghost
Copy link
Author

ghost commented Dec 21, 2021

/verify with intel/llvm#5148

@ghost ghost marked this pull request as ready for review December 21, 2021 16:35
@ghost ghost requested a review from AGindinson as a code owner December 21, 2021 16:35
@ghost ghost requested a review from vladimirlaz December 21, 2021 16:35
@ghost
Copy link
Author

ghost commented Dec 21, 2021

A test failed against the intel/llvm#5148 PR but this test is not related to the introduced changes. The PR can be merged after intel/llvm#5148

@bader
Copy link

bader commented Dec 22, 2021

@AGindinson, ping.

AGindinson
AGindinson previously approved these changes Dec 22, 2021
@ghost
Copy link
Author

ghost commented Dec 22, 2021

I'm waiting for the ongoing intel/llvm nightly-build update to trigger the checks against the compiler with the implemented feature.

@bader
Copy link

bader commented Dec 22, 2021

I'm waiting for the ongoing intel/llvm nightly-build update to trigger the checks against the compiler with the implemented feature.

I think Jenkins/llvm-test-suite checked against the compile with the implemented feature.
I see that SYCL :: DiscardEvents/invalid_event.cpp failed there. I'm not sure whether it's related to your patch, but if it's not, please, file a bug report and assign to @alexanderfle (he added this test recently).

@ghost
Copy link
Author

ghost commented Dec 22, 2021

@bader Thank you for the information, I've trigger the checks. I believe the issue with SYCL :: DiscardEvents/invalid_event.cpp is not related to the feature, also I see the issue occurs sporadically, for example it didn't occur in the checks for the PR in the intel/llvm repository. If it occurred, I would fill a bug report, no problem.

@bader
Copy link

bader commented Dec 22, 2021

It's already occurred, so please, file a bug regardless of the status of new run.

@ghost
Copy link
Author

ghost commented Dec 22, 2021

@bader I've opened the issue. I see the checks fail since they don't see latest changes in the sycl runtime/headers. I'm going to re-trigger the checks tomorrow.

@bader
Copy link

bader commented Dec 22, 2021

Okay. Follow-up on this with @vladimirlaz.

Pavel Samolysov added 6 commits December 23, 2021 10:26
…tomic_ref

Signed-off-by: Pavel Samolysov <pavel.samolysov@intel.com>
Signed-off-by: Pavel Samolysov <pavel.samolysov@intel.com>
Signed-off-by: Pavel Samolysov <pavel.samolysov@intel.com>
Signed-off-by: Pavel Samolysov <pavel.samolysov@intel.com>
Unfortunately, the CUDA backend in DPCPP has had no generic address space
support yet. The tests for the generic address space must fail on the CUDA
backend. To let us mark these tests as expectedly failed, the tests should
be moved to separated .cpp files.

Signed-off-by: Pavel Samolysov <pavel.samolysov@intel.com>
Signed-off-by: Pavel Samolysov <pavel.samolysov@intel.com>
@ghost ghost requested a review from AGindinson December 23, 2021 07:26
Copy link

@AGindinson AGindinson left a comment

Choose a reason for hiding this comment

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

FYI, the precommit tests can be retriggered via the Jenkins UI - unless, of course, the re-run requires a later version of test sources to be rebased upon.

@ghost
Copy link
Author

ghost commented Dec 23, 2021

@AGindinson thank you for the information.

@ghost
Copy link
Author

ghost commented Dec 23, 2021

The failed tests aren't related to the introduced changes.

@bader bader merged commit 20117af into intel:intel Dec 23, 2021
myler pushed a commit to myler/llvm-test-suite that referenced this pull request Apr 12, 2022
- use @line macro;
 - print pointers in hexadecimal format.
aelovikov-intel pushed a commit to aelovikov-intel/llvm that referenced this pull request Mar 27, 2023
…tomic_ref (intel/llvm-test-suite#619)

Unfortunately, the CUDA backend in DPCPP has had no generic address space
support yet. The tests for the generic address space must fail on the CUDA
backend. To let us mark these tests as expectedly failed, the tests should
be moved to separated .cpp files.

Signed-off-by: Pavel Samolysov <pavel.samolysov@intel.com>
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants