-
Notifications
You must be signed in to change notification settings - Fork 787
[SYCL][UR] Update UR loader version #8979
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
@@ -4,7 +4,7 @@ if (NOT DEFINED UNIFIED_RUNTIME_LIBRARY OR NOT DEFINED UNIFIED_RUNTIME_INCLUDE_D | |||
include(FetchContent) | |||
|
|||
set(UNIFIED_RUNTIME_REPO "https://github.com/oneapi-src/unified-runtime.git") | |||
set(UNIFIED_RUNTIME_TAG d6af758779db6eebdc419fd5e249302f566eb5de) | |||
set(UNIFIED_RUNTIME_TAG 7d22f806c45872dd6326a123915fe3c75be25b0e) |
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 think we should start having the discipline of updating only stable tags from unified runtime. So in this case, the next weekly tag containing this change. That way, we have always at least a standard practice of what we want to update and we know the changes we are bringing from unified-runtime are stable.
However, I understand the urgency of this fix.
@smaslov-intel : what do you think?
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.
Yes we should define validation+promotion process for accepting new UR versions.
Currently I do not think there is any additional validation being done on UR weekly tags. So they are not necessarily more stable than other commits.
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 agree we need improved stability guarantees going forward, what can that UR standalone testing be?
Of course, updating SYCL to use newer UR will additionally run pre-commit SYCL testing.
@@ -86,7 +86,6 @@ UR_DLLEXPORT ur_result_t UR_APICALL urGetEnqueueProcAddrTable( | |||
pDdiTable->pfnMemUnmap = nullptr; | |||
pDdiTable->pfnUSMMemcpy = nullptr; | |||
pDdiTable->pfnUSMPrefetch = nullptr; | |||
pDdiTable->pfnUSMMemAdvise = nullptr; |
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 was this removed? this is still on UR
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 was renamed, this variable is not in table anymore
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.
@bmyates could you add the new name, instead of removing it?
ur_pfnEnqueueUSMPrefetch_t pfnUSMPrefetch;
ur_pfnEnqueueUSMAdvise_t pfnUSMAdvise;
ur_pfnEnqueueUSMFill2D_t pfnUSMFill2D;
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.
added
@bader : it looks like this PR did not run any functional testing, did it? |
SYCL end-to-end testing is missing. Please, read this discussion to understand the root cause and suggested solutions. |
I see. @bmyates : please re-base |
Signed-off-by: Brandon Yates <brandon.yates@intel.com>
Signed-off-by: Brandon Yates <brandon.yates@intel.com>
Signed-off-by: Brandon Yates <brandon.yates@intel.com>
@smaslov-intel I rebased onto head of sycl |
This is ready to merge. Only pre-commit failure is unrelated to this PR and I see other PRs have the same test failing:
|
RHEL build fails in CI - http://icl-jenkins2.sc.intel.com:8080/blue/organizations/jenkins/SYCL_CI%2Fintel%2FBuild_PR_RHEL/detail/Build_PR_RHEL/5489/pipeline/152 |
Signed-off-by: Brandon Yates <brandon.yates@intel.com>
I think RHEL issue should be fixed now. pre-merge still failing due to unrelated #9008 |
rebased in #9053 |
Signed-off-by: Brandon Yates brandon.yates@intel.com