Conversation
It's downstream bug (one could argue with cuda-tile) when someone (not me) tries to intercepts the built tile-ir bytecode and uses cuda-tile-translate to convert it back to text, instruments it, then compiles it again. Only then they'll notice that it won't compile. That's because they would have a tile-ir kernel with a `load_ptr_tko` operation with no mask but a padding (which seems like something cutile-python allows). When this bytecode gets converted to text, since the padding (and mask) arguments are not keyworded and optional, it ends up generating tile-ir text in which the load_ptr_tko has the padding value inplace of the mask value. I'll mention this issue downstream, but I thought maybe it should be addressed here? unless there is a usage somehow for load_ptr_tko with a padding value and no mask value. Signed-off-by: Mohammad Issa <mo.issa.ok@gmail.com>
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
It's downstream bug (one could argue with cuda-tile) when someone (not me) tries to intercepts the built tile-ir bytecode and uses cuda-tile-translate to convert it back to text, instruments it, then compiles it again. Only then they'll notice that it won't compile.
That's because they would have a tile-ir kernel with a
load_ptr_tkooperation with no mask but a padding (which seems like something cutile-python allows). When this bytecode gets converted to text, since the padding (and mask) arguments are not keyworded and optional, it ends up generating tile-ir text in which the load_ptr_tko has the padding value inplace of the mask value.I'll mention this issue downstream, but I thought maybe it should be addressed here? unless there is a usage somehow for load_ptr_tko with a padding value and no mask value.
Description
Just that when generating the bytecode of a
load_ptr_tko, before assinging the value for the padding optional parameter, we check if a mask was given to begin with, if none were given, then the padding value is also set to none.Checklist