-
Notifications
You must be signed in to change notification settings - Fork 809
[SYCL][UR] Combined CTS fixes for CL adapter. #11794
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
[SYCL][UR] Combined CTS fixes for CL adapter. #11794
Conversation
d098f70 to
ab321fb
Compare
ab321fb to
76a4777
Compare
76a4777 to
d9eefd5
Compare
10bf762 to
7ddce35
Compare
kbenzie
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.
Not that oneapi-src/unified-runtime#1044 has merged, please pull in the latest sycl branch changes and update the UNIFIED_RUNTIME_REPO/UNIFIED_RUNTIME_TAG values as suggested. Then make this PR ready to review.
Pulls in fixes from oneapi-src/unified-runtime#1044 New UR commit hash 192e940 Also correct type in a couple of NUM_DEVICES PI queries, and implement a workaround in pi2ur for type mismatch between PI kernel num args query and UR equivalent.
81ff919 to
4bfc140
Compare
kbenzie
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.
UR LGTM
|
@intel/llvm-reviewers-runtime please review, it is intended for the next release. |
| case PI_KERNEL_INFO_NUM_ARGS: { | ||
| UrParamName = UR_KERNEL_INFO_NUM_ARGS; | ||
| break; | ||
| size_t NumArgs = 0; |
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 use use different types for different Num*?
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 turned out this particular query has always been size_t in UR, one of the fixes we're pulling in causes that to be properly validated now so we need to do this bit of translation. We do plan to fix this for the next release cycle oneapi-src/unified-runtime#1038
Pulls in fixes from oneapi-src/unified-runtime#1044