Optimize list constructions in micromega #10745
Closed
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
We prevent reallocating a lot of lists by threading the result of a fold function inside an accumulator. Includes the Micromega bootstrap.
This part of the code has no theorems about it, so it is more perilous to to preserve the same behaviour as before. Accumulator logic is bad for your brain. Also includes a Micromega bootstrap.
|
Bench shows nothing interesting. I guess nobody is using Lia in extreme cases. |
|
bedrock2 is missing in the benchmark, I guess that's because at the time this bench was run, our build was failing, but now it should work again, according to the ticks of our own CI. But I think bedrock2's current code would not run faster, because we avoid using lia in these extreme cases. @andres-erbsen built his own abstract interpretation based on intervals for this, but it would really be great to be able to call lia without having to worry wether it will run fast enough or not. And in fiat-crypto, if I understand correctly (/cc @JasonGross), it would simplify the proofs scripts if they could run So to summarize, I guess the reason nobody is using Lia in extreme cases is that Lia doesn't work in extreme cases at the moment. |
|
I have a hard time rebasing this after the merge of the zify PR. I might try reimplementing it from scratch. |
vbgl
added
the
needs: rebase
pi8027
added
the
part: micromega
|
Closing the PR in the meantime. |
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Outdated suggestions cannot be applied.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.
This PR reduces the number of list allocations in Micromega by using accumulators. On the example from #10743, the failure time is divided by almost 2, from ~18.5s to ~10.5s. There is probably a way to optimize even better, but for the time being this looked like a nice enhancement.
Bench running here.