[CK_TILE] fix enforcing fixed vectorsizes for ck tile conv #3344
+12
−2
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
jakpiase
requested review from
ThomasNing,
afagaj,
andriy-ca,
aosewski,
asleepzzz,
bartekxk,
carlushuang,
cgmillette,
coderfeli,
geyyer,
illsilin,
poyenc,
qianfengz,
shumway,
tenpercent and
vidyasagar-amd
as code owners
bartekxk
approved these changes
Dec 3, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 PR fixes a bug of inconsistent behavior of
FixedVectorSizeparameter inside CK Tile Convolutions and was discovered while porting convolution instances from old ck to ck tile. This parameter is used to disable automatic vectorsize selection and allow kernel user to specify some arbitral size of vectorloads.Current behavior is that:
gemm_pipeline_problem.hppusesFixedVectorSize, but we were lacking the enforcingifstatement insideuniversal gemm policy, which caused some part of the code to use automatic vectorload size selection, and the other part of code to use the arbitrary vectorload size, which was causing undefined behavior of some kernels and compilation errors for others, since the tile shapes were inconsistent. I have checked the behavior of this change on ckTileProfiler branch and it works for all cases tested there, but we definitely should create some tests to allow early detection and avoid that kind of bugs in the future.