Fix tilizing in tt-metal backend #462
Merged
Conversation
This file contains 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
rpavlovicTT
force-pushed
the
rpavlovic/tilizing_fix2
branch
from
63396fe to
91b905f
Compare
nsmithtt
approved these changes
Aug 21, 2024
There was a problem hiding this comment.
@rpavlovicTT, good catch and looks great! Once we fix the rank N support let's ship it.
Until now compiler would create single tilize block call with total amount of tiles needed to be generated. LLK actually needs to work on a single row of tiles at the time in order to know how much to stride. This fix changes lowering to TTMetal dialect and creates correct number of tilize/untilize calls. In between them it pops/pushes from/to given CBs. With SCF dialect we'll be able to insert loops, but for now these calls will be unrolled.
rpavlovicTT
force-pushed
the
rpavlovic/tilizing_fix2
branch
from
91b905f to
160334c
Compare
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Outdated suggestions cannot be applied.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.
Until now compiler would create single tilize block call with total amount of tiles needed to be generated. LLK actually needs to work on a single row of tiles at the time in order to know how much to stride.
This fix changes lowering to TTMetal dialect and creates correct number of tilize/untilize calls. In between them it pops/pushes from/to given CBs.
With SCF dialect we'll be able to insert loops, but for now these calls will be unrolled.