[SYCL] wait_and_throw invokes asynchronous handlers linked to the events that the event depends on.
#12202
Conversation
mmoadeli
temporarily deployed
to
WindowsCILock
GitHub Actions
Inactive
mmoadeli
changed the title
wait_and_throw to invoke asynchronous handlers linked to the events that the occurrence of an event relies on.
mmoadeli
changed the title
wait_and_throw to invoke asynchronous handlers linked to the events that the occurrence of an event relies on.wait_and_throw to invoke asynchronous handlers linked to the events that an event depends on.
mmoadeli
changed the title
wait_and_throw to invoke asynchronous handlers linked to the events that an event depends on.wait_and_throw invokes asynchronous handlers linked to the events that an event depends on.
mmoadeli
temporarily deployed
to
WindowsCILock
GitHub Actions
Inactive
mmoadeli
force-pushed
the
dependent-async-handler
branch
from
d0ad857
to
764f4e9
Compare
…this event relies on.
mmoadeli
force-pushed
the
dependent-async-handler
branch
from
764f4e9
to
bf9a906
Compare
mmoadeli
temporarily deployed
to
WindowsCILock
GitHub Actions
Inactive
mmoadeli
temporarily deployed
to
WindowsCILock
GitHub Actions
Inactive
mmoadeli
changed the title
wait_and_throw invokes asynchronous handlers linked to the events that an event depends on.wait_and_throw invokes asynchronous handlers linked to the events that the event depends on.
…requirements between two queues.
|
@dm-vodopyanov a friendly request for review please. |
|
@dm-vodopyanov, @Naghasan a friendly request for review please. |
|
@mmoadeli, the PR doesn't require any extra approvals, since @steffenlarsen is a member of runtime-reviewers team. However, I'm a bit hesitant to merge it right now considering that the spec PR is still open (even though it is approved). Tagging @gmlueck as one of the reviewers for extra inputs. Besides that, I wonder if this should be considered as a breaking change and put under |
thanks @AlexeySachkov |
|
Since Greg is OOO, I am tagging @Pennycook and @AerialMantis on this. It seems to me like the spec clarification has 2 approvals, so I figure it will get in soon. As to whether this is a breaking change is a good question though. |
|
|
||
| int main() { | ||
| { | ||
| test_exception_handler teh1; |
There was a problem hiding this comment.
I think it could be worth also having a test for the case that @gmlueck highlighted in the spec PR where there is no dependency chain between the two kernels but they are enqueued to the same queue.
It was decided that in this case we should all errors caught by the queue, so if there is a call to wait on the queue then both exceptions should be thrown.
…between the two kernels but they are enqueued to the same queue.
mmoadeli
had a problem deploying
to
WindowsCILock
GitHub Actions
Failure
|
This pull request is stale because it has been open 180 days with no activity. Remove stale label or comment or this will be automatically closed in 30 days. |
|
@AerialMantis It seems the corresponding spec PR is merged. Could this be merged in? |
Modifies
wait_and_throwto invoke asynchronous handlers linked to the events that an event depends on.There is a corresponding PR to clarify specification.