chore(l2): use SP1 patched crate for ecmul precompile #5133
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
Lines of code reportTotal lines added: Detailed view |
Benchmark Results ComparisonNo significant difference was registered for any benchmark run. Detailed ResultsBenchmark Results: BubbleSort
Benchmark Results: ERC20Approval
Benchmark Results: ERC20Mint
Benchmark Results: ERC20Transfer
Benchmark Results: Factorial
Benchmark Results: FactorialRecursive
Benchmark Results: Fibonacci
Benchmark Results: FibonacciRecursive
Benchmark Results: ManyHashes
Benchmark Results: MstoreBench
Benchmark Results: Push
Benchmark Results: SstoreBench_no_opt
|
Contributor
There was a problem hiding this comment.
Pull Request Overview
This PR refactors BN254 elliptic curve precompile functions (ecmul and ecpairing) by extracting common parsing and validation logic into reusable helper functions. The changes improve code organization and maintainability by separating concerns.
Key Changes:
- Extracted BN254 point parsing logic into
parse_bn254_g1,parse_bn254_g2, andparse_bn254_scalarhelper functions - Split validation logic from
validate_bn254_coordsinto separatevalidate_bn254_g1_coordsandvalidate_bn254_g2_coordsfunctions - Added SP1 feature-gated implementation of
bn254_g1_mulusing substrate_bn library
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
ilitteri
requested changes
Oct 31, 2025
Contributor
ilitteri
left a comment
ilitteri
left a comment
There was a problem hiding this comment.
The implementation looks good. I'll run some replays to ensure it works properly. Left some comments.
Comment on lines
-819
to
-888
| #[expect(unsafe_code, reason = "chunks_exact ensures the conversion is valid")] | ||
| let input: [u8; 192] = unsafe { input.try_into().unwrap_unchecked() }; | ||
| let (g1, g2) = parse_bn254_coords(&input); | ||
| if validate_bn254_coords(&g1, &g2)? { | ||
| batch.push((g1, g2)); | ||
| } |
Contributor
There was a problem hiding this comment.
Why do we remove this unsafe? We should double-check with @edg-l, @azteca1998, or @Oppen if this is not counterproductive.
Contributor
There was a problem hiding this comment.
Can't we keep this behaviour when neither sp1 nor risc0 features are enabled?
Contributor
Author
There was a problem hiding this comment.
The function to parse bn254 points was changed to take a slice instead of a reference to a fixed size array (so it can be reused for precompiles whose input have different sizes), so this unsafe is not necessary anymore.
The branching on the error case (which should never happen) could signify a perf. hit, I could change this to an unreachable_unchecked!() and it should be equivalent to before the change, but I'm not sure how impactful that can be (consider that we have other, more evident optimization opportunities, like not cloning and slicing bytes many times while parsing and converting to the crypto primitive types)
ilitteri
approved these changes
Oct 31, 2025
pablodeymo
approved these changes
Nov 4, 2025
xqft
added a commit
that referenced
this pull request
Nov 11, 2025
**Motivation** The ecmul precompile represents 10% of the total proving cycles, and we are not using the corresponding SP1 patched crate for this operation that uses a zkVM precompile. This PR fixes that and reduces 81k cycles (86% of ecmul, 10% of total) **Flamegraphs** before: <img width="2936" height="1134" alt="image" src="https://github.com/user-attachments/assets/42212408-8f2e-4dd4-80ea-c9b87a8383ef" /> after: <img width="2716" height="1220" alt="image" src="https://github.com/user-attachments/assets/f82e25f0-a7cb-43e9-ae4d-307a39bd9d0a" /> **Proving times** Mainnet blocks proven in a RTX 4090 | Block number | main | l2/ecmul_sp1 | |---------------|---------|----------------| | 23426993 | 13m 19s | 13m 20s | | 23426994 | 08m 42s | 08m 05s | | 23426995 | 06m 39s | 06m 35s | | 23426996 | 11m 24s | 10m 50s | --------- Co-authored-by: Copilot <[email protected]> Co-authored-by: Ivan Litteri <[email protected]> Co-authored-by: ilitteri <[email protected]>
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.
Motivation
The ecmul precompile represents 10% of the total proving cycles, and we are not using the corresponding SP1 patched crate for this operation that uses a zkVM precompile.
This PR fixes that and reduces 81k cycles (86% of ecmul, 10% of total)
Flamegraphs
before:
after:
Proving times
Mainnet blocks proven in a RTX 4090