IPIN Cost Multiplier #2843
Merged
IPIN Cost Multiplier #2843
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
|
Thanks @amin1377 . I don't expect this will lead to any meaningful change, but could you run a QoR check on Titan to be sure? |
|
Titan quick qor: |
|
@vaughnbetz: The QoR remains nearly the same, but placement time has increased by 6%, likely contributing to a 3% rise in overall runtime. I ran the benchmark on SAVI to minimize machine load fluctuations, but it's possible the load changed during the test with the new code. I don’t think the changes made in this PR would significantly impact runtime. |
|
Thanks. Yes it's hard to imagine how this could change place time |
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Outdated suggestions cannot be applied.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.
The IPIN cost multiplier was set to a value below one to prioritize pulling IPINs faster during wave expansion from the heap, enabling quicker access to the SINK if the IPIN is connected to the target SINK. Previously, this multiplier was set to 0.95, which is not perfectly representable in the IEEE 754 floating-point format. To ensure consistent results across all platforms, a condition was added to use 0.875 (1/2 + 1/4 + 1/8) when the SPEC flag was enabled. However, since 0.95 and 0.875 are very close in value, we have now removed the conditional logic and only used 0.875.