-
Notifications
You must be signed in to change notification settings - Fork 794
[SYCL][L0] Enable global offset support for level_zero. #3593
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We need to cache extensions such that the search is not repeated on each kernel launch
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
can we further update some local "static bool" value such that the map lookup is performed only once?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
sure, that can be done during the initialization. In that case, defining this generic helper function findDriverExtension() won't be necessary anymore in this PR
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why do you think we need to give error in this case?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
would you remove this?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I kept it as it was in Gail's patch. My reason for keeping it is to do an early return and saving the overhead of the next zeDriverGetExtensionProperties call.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
it made sense when it was looking for "global-offset" specifically" but not when it is made a general query
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
removing it.
Signed-off-by: rbegam <[email protected]>
Signed-off-by: rbegam <[email protected]>
Signed-off-by: rbegam <[email protected]>
Signed-off-by: rbegam <[email protected]>
Signed-off-by: rbegam <[email protected]>
|
ping @smaslov-intel |
Signed-off-by: rbegam <[email protected]>
smaslov-intel
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
|
The pre-commit fail is expected as with this PR the corresponding test (prev marked as XFAIL) is now passing: Unexpectedly Passed Tests (1): The test needs to be enabled which is done in this PR: intel/llvm-test-suite#261 Local testing shows the test passes with this compiler change and vice versa. |
|
Please do not merge the PRs. I have restarted testing to make sure that no regressions introduced. |
Please, let me know when you think we can merge it. I'm also waiting for @romanovvlad to confirm that his comments are addressed. |
|
Now it is OK to merge. Tests have passed |
romanovvlad
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Approving to indicate that my minor comment is resolved. Have not reviewed entire patch.
Signed-off-by: rbegam [email protected]