[WIP][Place] Fixed Issue Where Macros Were Not Being Created - #3519
Open
AlexandreSinger wants to merge 2 commits into
Open
[WIP][Place] Fixed Issue Where Macros Were Not Being Created#3519AlexandreSinger wants to merge 2 commits into
AlexandreSinger wants to merge 2 commits into
Conversation
For special architectures, there may exist direct connections from one sub-tile to another within the same tile. For blocks which use these direct connections, they must be placed relative to specific constraints. Normally, to enforce these constraints during packing we use macros; however macros were not being created for this case. Although the tests were passing, they were passing due to random chance (since the initial placer happened to place them in the correct order and the annealer did not change them). Now that AP is coming in, this is actually causing an issue. Fixed the macro creation code so it actually creates a macro for this case.
github-actions
Bot
added
VPR
AlexandreSinger
changed the title
vaughnbetz
reviewed
Jun 2, 2026
|
|
||
| // Pure-source macro detection: | ||
| // Handles blocks that have DRIVER pins for a direct but no RECEIVER pins for that | ||
| // same direct (e.g. a multiplier driving memory blocks via sub-tile direct connections). |
Contributor
There was a problem hiding this comment.
- This is a very long routine. Suggest adding some helper functions to break it up.
- It's not clear to me what a 'pure-source' macro is (maybe spans two different types of blocks?), or what the overall algorithm is fo identifying them. I suggest definiing the term, and giving an overview of the algorithm here,
Contributor
Author
There was a problem hiding this comment.
I agree, this was the PR I was discussing with you a week or two ago. It fixes the issue of the inter-sub-tile macros; however I am not confident in the quality of the solution (Claude had a heavy hand in creating it). I left this PR open for @amin1377 to look at if he was curious.
I did verify that it fixes the issue without breaking the rest of VPR (as far as I can tell); but the code is a bit magical.
Contributor
There was a problem hiding this comment.
Ok, to @amin1377 to see if he can make a better solution.
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.
For special architectures, there may exist direct connections from one sub-tile to another within the same tile. For blocks which use these direct connections, they must be placed relative to specific constraints.
Normally, to enforce these constraints during packing we use macros; however macros were not being created for this case. Although the tests were passing, they were passing due to random chance (since the initial placer happened to place them in the correct order and the annealer did not change them). Now that AP is coming in, this is actually causing an issue.
Fixed the macro creation code so it actually creates a macro for this case.