fixing backport #551
Merged
fixing backport #551
Conversation
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
This PR fixes several issues related to fixpoint rewrites:
1. A subtle bug in `Pass.fixpoint` lead to rewrites being applied for
`max_iter` iterations, even if they had already converged.
2. Several compiler passes were incorrectly returning
`has_done_something=True`, even if they had not done anything, leading
to maxed out fixpoint loops.
3. The CSE rewrite would prematurely exit, leading to more fixpoint
iterations.
I checked that all tests in the kirin test suite lead to fixpoint
rewrites that converge before reaching `max_iter`. It's still possible
that there are more rewrite passes that need to be fixed. We should
consider raising an error or at least a warning if a fixpoint pass does
not converge, as this will immediately point us to any such issues in
the future.
**The default `max_iter` is set to 32. With this fix, many passes now
iterate only once or twice instead of 32 times. So in many cases there
is a 32x speedup. Some passes also contain a fixpoint within a fixpoint.
They can now become 1000x faster.**
I noticed these issues when running the following script
```python
@qasm2.extended # O(n) scaling
def main_fun():
qreg = qasm2.qreg(1)
for _ in range(n):
qasm2.h(qreg[0])
QASM2().emit(main_fun) # ~O(n^1.5)
```
This script runs really slow for large `n`. But even more important,
recording the duration of compilation and QASM2 emission, I found
non-linear scaling ~ O(n^1.5). Any compiler pass must be linear in the
program size (or at worst n log n).
<img width="331" height="250" alt="image"
src="https://github.com/user-attachments/assets/c914db82-e2c6-4821-8a0e-b00ab19b432c"
/>
With the fixes in this PR, the execution time of this script is
significantly faster and the computational complexity is back to the
expected O(n):
<img width="331" height="245" alt="image"
src="https://github.com/user-attachments/assets/35480c95-29b5-4117-b796-c4cc891d1cb0"
/>
These issues also affect the runtime of @cduck's flair-to-squin pipeline
- but I think there are additional slowdowns that I'm still looking
into.
Contributor
|
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.
No description provided.