Skip to content

Conversation

@hero78119
Copy link
Collaborator

@hero78119 hero78119 commented Jan 28, 2026

Related to #1233
refer https://eprint.iacr.org/2024/108.pdf "sec 3.1"

benchmark

send one less interpolation show significant improvement on prove_tower_relation_gpu. For prove_main_constraints, pcs_opening there is no significant different and most likely noise

Operation master (s) degree-1 (s) Improve (master→degree-1)
prove_tower_relation_gpu 32.689 26.446 -19.14%
prove_main_constraints 9.932 10.146 +2.15%
pcs_opening 19.081 19.386 +1.60%
Layer master (s) degree-1 (s) Improve (master→degree-1)
app_prove 87.000 82.100 5.63%

raw data

after degree - 1 now trigger more gap, we need to optimize per shard step records generation + witness assignment

}
}
let final_sum = SepticPoint::from_affine(final_sum_x, final_sum_y);
assert_eq!(zerocheck_proof.extract_sum(), E::ZERO);
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

this util function extract_sum no longer valid as now uni-variate lack of p0. We can do the zero check in sanity stage

Copy link
Collaborator

@kunxian-xia kunxian-xia left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM 🎉

@kunxian-xia kunxian-xia added this pull request to the merge queue Feb 2, 2026
Merged via the queue into master with commit 3f5214f Feb 2, 2026
5 checks passed
@kunxian-xia kunxian-xia deleted the feat/skip_eval_0 branch February 2, 2026 07:55
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants