Attribute compute_io_bound IO stalls to the starved input - #108
Merged
Conversation
compute_io_bound could say a layer was IO bound but not what it was waiting on. It now also programs the core tile one row above the existing one, whose memory module counts two combo events: the core's lock stall ANDed with S2MM channel 0 starvation, and the same ANDed with channel 1. Channel 0 carries the IFM and channel 1 the weights, so comparing the two names the input that starved the core. A combo event can only combine events of one module and the starvation events belong to the memory module, so the lock stall reaches it as an intra-tile broadcast, the same route the first tile already uses. The core tile metric set is also selected by name rather than by a boolean, which the second metric set this was originally written for needed and which remains the clearer interface. Co-authored-by: Cursor <cursoragent@cursor.com>
jvillarre
approved these changes
Aug 25, 2026
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
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
compute_io_bound could say a layer was IO bound but not what it was waiting on. It now also programs the core tile one row above the existing one, whose memory module counts two combo events: the core's lock stall ANDed with S2MM channel 0 starvation, and the same ANDed with channel 1. Channel 0 carries the IFM and channel 1 the weights, so comparing the two names the input that starved the core.
A combo event can only combine events of one module and the starvation events belong to the memory module, so the lock stall reaches it as an intra-tile broadcast, the same route the first tile already uses.
The core tile metric set is also selected by name rather than by a boolean, which the second metric set this was originally written for needed and which remains the clearer interface.