Skip to content

aiur: verify u32 addition from hinted sum - #556

Draft
gabriel-barrett wants to merge 2 commits into
mainfrom
perf/u32-add-hint
Draft

aiur: verify u32 addition from hinted sum#556
gabriel-barrett wants to merge 2 commits into
mainfrom
perf/u32-add-hint

Conversation

@gabriel-barrett

Copy link
Copy Markdown
Member

u32_add optimization

Codegen regenerated; all 71 kernel FFT pins and the shard-pipeline
aggregate re-measured: every pin drops (median -3.7%, up to -7.9% on
the string/byte-heavy checks), shard aggregate -9.9%. aiur (cargo +
prove suite incl. parity), ixvm, multi-stark and recursive-verifier
suites pass; fmt, clippy, deny clean.
@gabriel-barrett

Copy link
Copy Markdown
Member Author

!benchmark aiur-recursive fresh

@argument-ci-bot

argument-ci-bot Bot commented Aug 12, 2026

Copy link
Copy Markdown

!benchmark — main vs d2102ba

backends: aiur-recursive · envs: InitStd · set: primary · shard: 0 · baseline: fresh (base-SHA run, bencher bypassed)

aiur-recursive · InitStd — main from: base run @ 5392d37 (fresh — bencher bypassed)

1 constant · 1 with regressions · 1 with improvements (|Δ| > 3.0% on any metric).

constant recursive-prove-time (main) recursive-prove-time (PR) Δ% recursive-peak-ram (main) recursive-peak-ram (PR) Δ% recursive-proof-size (main) recursive-proof-size (PR) Δ% recursive-verify-time (main) recursive-verify-time (PR) Δ% recursive-execute-time (main) recursive-execute-time (PR) Δ% recursive-fft-cost (main) recursive-fft-cost (PR) Δ% prove-time (main) prove-time (PR) Δ% proof-size (main) proof-size (PR) Δ% verify-time (main) verify-time (PR) Δ% peak-ram (main) peak-ram (PR) Δ%
Nat.add_comm 26.504 s 31.599 s +19.2% (1.19× slower) ⚠️ 80.81 GiB 64.22 GiB -20.5% (1.26× smaller) 🟢 5.48 MiB 5.36 MiB -2.2% 33.1 ms 33.8 ms +2.2% 4.142 s 12.498 s +201.8% (3.02× slower) ⚠️ 148.08B 122.99B -16.9% (1.20× fewer) 🟢 1.017 s 1.033 s +1.5% 9.11 MiB 8.99 MiB -1.3% 58.2 ms 57.9 ms -0.5% 3.44 GiB 4.22 GiB +22.6% (1.23× larger) ⚠️

Workflow logs

Codegen regenerated; all 71 kernel FFT pins and the shard-pipeline
aggregate re-measured: every pin drops again (median -1.5%, up to
-3.4% on the string/byte-heavy checks), shard aggregate -4.3% on top
of the hinted-adder commit (-13.8% combined). aiur (cargo + prove
suite incl. parity), ixvm, multi-stark and recursive-verifier suites
pass; fmt, clippy, deny clean.
@gabriel-barrett

Copy link
Copy Markdown
Member Author

!benchmark aiur-recursive fresh

@argument-ci-bot

argument-ci-bot Bot commented Aug 12, 2026

Copy link
Copy Markdown

!benchmark — main vs 7257a73

backends: aiur-recursive · envs: InitStd · set: primary · shard: 0 · baseline: fresh (base-SHA run, bencher bypassed)

aiur-recursive · InitStd — main from: base run @ 5392d37 (fresh — bencher bypassed)

1 constant · 1 with regressions · 1 with improvements (|Δ| > 3.0% on any metric).

constant recursive-prove-time (main) recursive-prove-time (PR) Δ% recursive-peak-ram (main) recursive-peak-ram (PR) Δ% recursive-proof-size (main) recursive-proof-size (PR) Δ% recursive-verify-time (main) recursive-verify-time (PR) Δ% recursive-execute-time (main) recursive-execute-time (PR) Δ% recursive-fft-cost (main) recursive-fft-cost (PR) Δ% prove-time (main) prove-time (PR) Δ% proof-size (main) proof-size (PR) Δ% verify-time (main) verify-time (PR) Δ% peak-ram (main) peak-ram (PR) Δ%
Nat.add_comm 26.528 s 32.516 s +22.6% (1.23× slower) ⚠️ 82.19 GiB 59.44 GiB -27.7% (1.38× smaller) 🟢 5.48 MiB 5.31 MiB -3.0% 🟢 33.1 ms 32.8 ms -1.1% 4.142 s 14.644 s +253.6% (3.54× slower) ⚠️ 148.08B 112.77B -23.9% (1.31× fewer) 🟢 1.010 s 1.057 s +4.6% ⚠️ 9.11 MiB 8.94 MiB -1.8% 57.5 ms 56.5 ms -1.7% 4.11 GiB 4.44 GiB +8.2% (1.08× larger) ⚠️

Workflow logs

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.

1 participant