-
Notifications
You must be signed in to change notification settings - Fork 131
[SYCL] Update tests with -fno-sycl-id-queries-fit-in-int option #220
Conversation
The changes are due to the PR: intel/llvm#3427 Signed-off-by: Soumi Manna <[email protected]>
This reverts commit 258eec3.
The changes are due to the PR: intel/llvm#3427 Signed-off-by: Soumi Manna <[email protected]>
The failures seem unrelated to my change. |
@@ -1,4 +1,4 @@ | |||
// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out | |||
// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple -fno-sycl-id-queries-fit-in-int %s -o %t.out |
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 wonder, why those changes are needed at all?
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.
Thanks @AlexeySachkov for reviewing the patch.
intel/llvm#3427: Enables the option -fsycl-id-queries-fit-in-int to true by default. sub_group_as.cpp passes when the option is false on default mode but fails with default ON. So I added -fno-sycl-id-queries-fit-in-int to disable the option and match with the old behavior.
sycl-id-queries-fit-in-int makes a big impact on performance for DPC++ in many cases. The test failures currently block intel/llvm#3427. Is there any workaround i can use here? Thanks.
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 looks like root cause for failing test is the same with post-commit failures after merging intel/llvm#3427
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.
@vladimirlaz, PR: intel/llvm#3427: which has not been merged yet. I am waiting for this PR to be merged first before merging PR: intel/llvm#3427
@vladimirlaz, I have created intel/llvm#3495 for the regression and assigned to you. Thanks. |
@@vladimirlaz, Any update on this? Thanks. Blocks PR: intel/llvm#3427 |
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'm okay to submit this patch since we have an issue filed to fix the regression
Thanks @AlexeySachkov for the review. @vladimirlaz, could you please commit it?. Thanks. |
…el/llvm-test-suite#220) Signed-off-by: Soumi Manna <[email protected]>
The changes are due to the PR: intel/llvm#3427
Signed-off-by: Soumi Manna [email protected]