Skip to content

New commands for profiling test#1780

Open
gwawiork wants to merge 1 commit intoKhronosGroup:mainfrom
gwawiork:profiling_new_cmds
Open

New commands for profiling test#1780
gwawiork wants to merge 1 commit intoKhronosGroup:mainfrom
gwawiork:profiling_new_cmds

Conversation

@gwawiork
Copy link
Copy Markdown
Contributor

@gwawiork gwawiork commented Jul 6, 2023

  • clEnqueueMarkerWithWaitList
  • clEnqueueBarrierWithWaitList
    The test verifies if:
  • sync commands have proper timestamps
  • ndrange commands before and after sync have proper timestamps (marker, barrier works correctly)

test_error(error, "Unable to set argument for test kernel");

error = create_single_kernel_helper(context, &program, &kernel3, 1,
&test_kernel, "test1");
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Should this be test3?

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

now there is one kernel in the code run multiple times

fn(queue_with_props, 3, &events_list_set2[0], &eventEnqueueMarkerSet2);
test_error(error, "Unable to enqueue sync command");

error = clWaitForEvents(1, &eventEnqueueMarkerSet2);
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Should clWaitForEvents be switched to eventEnqueueMarkerSet1 since that's what's being checked in the following lines, the command queue is out of order, and set 1 doesn't wait for set 2 to complete? Or maybe clFinish would be more readable?

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

added waiting on sync_event1 beside of waiting on sync_event2

@gwawiork
Copy link
Copy Markdown
Contributor Author

@lakshmih Thanks for review. Unfortunately I see some errors on our internal driver version. Until these issues will not be solved I do not plan any updates in this PR. I will let you know when it will be ready for review.

@lakshmih
Copy link
Copy Markdown
Contributor

@lakshmih Thanks for review. Unfortunately I see some errors on our internal driver version. Until these issues will not be solved I do not plan any updates in this PR. I will let you know when it will be ready for review.

Thank you. Could you review the focused review label and add it back when ready?

@gwawiork
Copy link
Copy Markdown
Contributor Author

@lakshmih Thanks for review. Unfortunately I see some errors on our internal driver version. Until these issues will not be solved I do not plan any updates in this PR. I will let you know when it will be ready for review.

Thank you. Could you review the focused review label and add it back when ready?

I think I do not have permission to modify labels.

@lakshmih
Copy link
Copy Markdown
Contributor

lakshmih commented Oct 2, 2023

@lakshmih Thanks for review. Unfortunately I see some errors on our internal driver version. Until these issues will not be solved I do not plan any updates in this PR. I will let you know when it will be ready for review.

Thank you. Could you review the focused review label and add it back when ready?

I think I do not have permission to modify labels.

Oh, thanks for letting me know. I removed it for now, we can add it back when ready.

WojciechKonior added a commit to WojciechKonior/OpenCL-CTS that referenced this pull request Apr 17, 2025
KhronosGroup#1780 pr remake.
Test verifies if clEnqueueMarkerWithWaitList and clEnqueueBarrierWithWaitList
sync commands have proper timestamps.

Signed-off-by: Wojciech Konior <wojciech.konior@intel.com>
WojciechKonior added a commit to WojciechKonior/OpenCL-CTS that referenced this pull request Apr 17, 2025
KhronosGroup#1780 pr remake.
Test verifies if clEnqueueMarkerWithWaitList and clEnqueueBarrierWithWaitList
sync commands have proper timestamps.

Signed-off-by: Wojciech Konior <wojciech.konior@intel.com>
WojciechKonior added a commit to WojciechKonior/OpenCL-CTS that referenced this pull request Apr 17, 2025
KhronosGroup#1780 pr remake.
Test verifies if clEnqueueMarkerWithWaitList and clEnqueueBarrierWithWaitList
sync commands have proper timestamps.

Signed-off-by: Wojciech Konior <wojciech.konior@intel.com>
WojciechKonior added a commit to WojciechKonior/OpenCL-CTS that referenced this pull request May 5, 2025
KhronosGroup#1780 pr remake.
Test verifies if clEnqueueMarkerWithWaitList and clEnqueueBarrierWithWaitList
sync commands have proper timestamps.

Signed-off-by: Wojciech Konior <wojciech.konior@intel.com>
WojciechKonior added a commit to WojciechKonior/OpenCL-CTS that referenced this pull request May 5, 2025
KhronosGroup#1780 pr remake.
Test verifies if clEnqueueMarkerWithWaitList and clEnqueueBarrierWithWaitList
sync commands have proper timestamps.

Signed-off-by: Wojciech Konior <wojciech.konior@intel.com>
WojciechKonior added a commit to WojciechKonior/OpenCL-CTS that referenced this pull request May 5, 2025
KhronosGroup#1780 pr remake.
Test verifies if clEnqueueMarkerWithWaitList and clEnqueueBarrierWithWaitList
sync commands have proper timestamps.

Signed-off-by: Wojciech Konior <wojciech.konior@intel.com>
@gwawiork gwawiork force-pushed the profiling_new_cmds branch 3 times, most recently from 9a0f837 to b01c09b Compare October 6, 2025 14:08
@gwawiork
Copy link
Copy Markdown
Contributor Author

gwawiork commented Oct 6, 2025

This PR is refactored and fixed. Ready for review
@lakshmih FYI
I think #2374 is not needed

@gwawiork gwawiork force-pushed the profiling_new_cmds branch from b01c09b to aa101f6 Compare October 7, 2025 10:16
Copy link
Copy Markdown
Contributor

@bashbaug bashbaug left a comment

Choose a reason for hiding this comment

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

Thanks, this is looking good to me, still needs a non-Intel reviewer.

@gwawiork gwawiork force-pushed the profiling_new_cmds branch from aa101f6 to 36c598d Compare March 30, 2026 16:28
@gwawiork gwawiork force-pushed the profiling_new_cmds branch from 36c598d to 37e08ae Compare March 31, 2026 10:55
Copy link
Copy Markdown
Contributor

@bashbaug bashbaug left a comment

Choose a reason for hiding this comment

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

Thanks! Latest changes LGTM. Needs a non-Intel reviewer.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants