[SYCL][FPGA] Allow safelen = 0 with ivdep attribute #6336
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This patch allows safelen = 0 with ivdep attribute and emits suppressible warning when safelen of 0 or 1 is used Current behaviour: ivdep with safelen of 1 is allowed with no warning. Safelen of 0 causes an error. Desired behaviour: safelen of 1 and 0 should both be allowed Both should emit a warning. The warning should be suppressible with a -Wno flag. Justification: The safelen parameter of an ivdep will sometimes be set based on a template argument. It makes this type of code much cleaner if the setting of 0 is permitted. The warning for both the 0 and 1 case is helpful for users who might misunderstand the meaning of the safelen parameter and set a value of 1 (or possibly 0) thinking it will have some effect. The warning should be suppressible for those developers using the templated coding pattern described above who understand and accept that a safelen of 0 will have no effect. Signed-off-by: Soumi Manna <soumi.manna@intel.com>
Signed-off-by: Soumi Manna <soumi.manna@intel.com>
Signed-off-by: Soumi Manna <soumi.manna@intel.com>
Contributor
premanandrao
left a comment
premanandrao
left a comment
There was a problem hiding this comment.
Do we need to modify the docs?
Signed-off-by: Soumi Manna <soumi.manna@intel.com>
Contributor
Author
Thanks @premanandrao for review. I have updated docs as well. |
erichkeane
reviewed
Jun 22, 2022
erichkeane
reviewed
Jun 22, 2022
Signed-off-by: Soumi Manna <soumi.manna@intel.com>
smanna12
requested review from
AaronBallman,
elizabethandrews,
erichkeane and
premanandrao
Signed-off-by: Soumi Manna <soumi.manna@intel.com>
AaronBallman
previously approved these changes
Jun 22, 2022
Contributor
AaronBallman
left a comment
AaronBallman
left a comment
There was a problem hiding this comment.
LGTM as far as that goes
Contributor
Author
Thanks @AaronBallman for reviews and feedbacks. |
Signed-off-by: Soumi Manna <soumi.manna@intel.com>
premanandrao
previously approved these changes
Jun 23, 2022
Signed-off-by: Soumi Manna <soumi.manna@intel.com>
elizabethandrews
approved these changes
Jun 23, 2022
Contributor
Author
|
Thank you for the reviews everyone. @intel/llvm-gatekeepers, this PR is ready for a merge. Thank you |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This patch allows safelen = 0 with ivdep attribute and emits suppressible warning when safelen of 0 or 1 is used
Current behaviour:
ivdep with safelen of 1 is allowed with no warning. Safelen of 0 causes an error.
Desired behaviour:
safelen of 1 and 0 should both be allowed but will have no effect on the loop.
Both should emit a warning. The warning should be suppressible with a -Wno flag.
Justification:
The safelen parameter of an ivdep will sometimes be set based on a template argument.
It makes this type of code much cleaner if the setting of 0 is permitted.
The warning for both the 0 and 1 case is helpful for users who might misunderstand the
meaning of the safelen parameter and set a value of 1 (or possibly 0) thinking it will have some effect.
The warning should be suppressible for those developers using the templated coding pattern described
above who understand and accept that a safelen of 0 will have no effect.
Signed-off-by: Soumi Manna soumi.manna@intel.com