[TKW] Propagate dim index in thread shape analysis #288
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
Signed-off-by: Ivan Butygin <[email protected]>
Signed-off-by: Ivan Butygin <[email protected]>
Signed-off-by: Ivan Butygin <[email protected]>
Signed-off-by: Ivan Butygin <[email protected]>
Signed-off-by: Ivan Butygin <[email protected]>
Hardcode84
changed the title
Hardcode84
changed the title
harsh-nod
requested changes
Nov 22, 2024
| @@ -1349,6 +1349,10 @@ def num_reduction_dims(self) -> int: | |||
| def reduction_dim(self) -> IndexSymbol: | |||
| return self.dim | |||
|
|
|||
| @property | |||
There was a problem hiding this comment.
Can you add a comment explaining why this is necessary?
| @@ -51,7 +57,7 @@ def set_index_size(custom: CustomOp, target_dim_sizes: list[DimSize]): | |||
| # Anchor Indicies and Conflict resolution helpers | |||
| ################################################################# | |||
|
|
|||
| anchorOpTypes = (Read, Write, MMA, ReduceOp, Reshape) | |||
| anchorOpTypes = (Read, Write, MMA, ReduceOp, Reshape, Permute) | |||
| noHandleTypes = (Placeholder, Output, ExtractSlice, Allocate) | |||
There was a problem hiding this comment.
Also a comment explaining why Permute is added as an anchor op?
raikonenfnu
reviewed
Nov 22, 2024
| @@ -51,7 +57,7 @@ def set_index_size(custom: CustomOp, target_dim_sizes: list[DimSize]): | |||
| # Anchor Indicies and Conflict resolution helpers | |||
| ################################################################# | |||
|
|
|||
| anchorOpTypes = (Read, Write, MMA, ReduceOp, Reshape) | |||
| anchorOpTypes = (Read, Write, MMA, ReduceOp, Reshape, Permute) | |||
There was a problem hiding this comment.
Can we add the forward propagation of permute, just as safety measure to ensure we won't be generating "valid" but incorrect IRs. Speaking from experience, would be much better for program to crash than to debug why MLIR is wrong and where the wrong is coming from. 😄
Signed-off-by: Ivan Butygin <[email protected]>
Signed-off-by: Ivan Butygin <[email protected]>
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.
Refactor
thread_shape_analysisto take into account entire index instead of just elements per thread count.