multi-stark: adopt flat quotient commitment (rev 678f8e1)

[gabriel-barrett]

Bump multi-stark to 678f8e1c96b3d67eaddaa9aa8d03cd7695f97081, which commits each circuit's quotient coefficient slices as one wide matrix on the trace domain (opened once at ζ) instead of one matrix per slice on the split sub-cosets.

Port the recursive verifier to the new proof shape:

• OOD: quotient(ζ) is now the power series Σᵢ ζ^(i·n)·cᵢ(ζ) over the slice values reconstructed from the single opened row, replacing the sub-domain Lagrange recombination (zps products over vanishing polynomials). Adds the check_shapes width assert (qd slices per row) and the one-matrix-per-circuit round-length check in verify().
• PCS: the quotient FRI batch is one matrix per circuit at the same per-circuit heights as the stages; the per-chunk open loop and the chunk-expanded height list are gone, and the verifying-key circuit list no longer threads through the PCS phase.
• Delete the dead zps machinery (g_pow, vanish_chunk, zps_prod, quotient_sum, log_qd_of, open_q_chunks, heights_quotient, repeat_g).

Regenerate crates/ixvm-codegen/src/aiur_multi_stark.rs (215 aiur fns).

Verified: multi-stark self-tests and the recursive-verifier end-to-end suite pass (honest proof accepted, codegen/interpreter parity holds, tampered proof and tampered claim rejected).

[gabriel-barrett]

!benchmark aiur fresh

[argument-ci-bot]

!benchmark — main vs c0b3b5a

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

aiur · InitStd · prove — main from: base run @ 6e29ac9 (fresh — bencher bypassed)

constant	prove-time (main)	prove-time (PR)	Δ%	throughput (const/s) (main)	throughput (const/s) (PR)	Δ%	peak-ram (main)	peak-ram (PR)	Δ%	execute-time (main)	execute-time (PR)	Δ%	verify-time (main)	verify-time (PR)	Δ%	proof-size (main)	proof-size (PR)	Δ%	fft-cost (main)	fft-cost (PR)	Δ%
Array.extract_append	51.102 s	48.467 s	-5.2% (1.05× faster) 🟢	33.380	35.200	+5.5% (1.05× faster) 🟢	96.75 GiB	96.31 GiB	-0.4%	10.992 s	10.423 s	-5.2% (1.05× faster) 🟢	249.9 ms	167.3 ms	-33.1% (1.49× faster) 🟢	26.35 MiB	25.88 MiB	-1.8%	36.21B	36.21B	+0.0%
ByteArray.utf8DecodeChar?_utf8EncodeChar_append	47.842 s	47.150 s	-1.4%	60.410	61.290	+1.5%	102.44 GiB	102.03 GiB	-0.4%	9.166 s	9.012 s	-1.7%	185.9 ms	168.6 ms	-9.3% (1.10× faster) 🟢	26.44 MiB	25.96 MiB	-1.8%	38.11B	38.11B	+0.0%
Char.ofOrdinal_le_of_le	40.899 s	37.834 s	-7.5% (1.08× faster) 🟢	69.980	75.650	+8.1% (1.08× faster) 🟢	80.23 GiB	79.88 GiB	-0.4%	6.722 s	6.751 s	+0.4%	260.5 ms	166.2 ms	-36.2% (1.57× faster) 🟢	26.38 MiB	25.90 MiB	-1.8%	29.15B	29.15B	+0.0%
Vector.extract_append._proof_2	27.717 s	27.717 s	+0.0%	51.990	51.990	+0.0%	53.31 GiB	53.12 GiB	-0.4%	5.761 s	5.694 s	-1.2%	175.2 ms	168.7 ms	-3.7% 🟢	26.02 MiB	25.55 MiB	-1.8%	21.18B	21.18B	+0.0%
String.split	22.263 s	23.328 s	+4.8% ⚠️	87.550	83.550	-4.6% ⚠️	48.21 GiB	48.06 GiB	-0.3%	3.080 s	3.153 s	+2.4%	174.0 ms	163.4 ms	-6.1% (1.06× faster) 🟢	26.39 MiB	25.92 MiB	-1.8%	14.22B	14.22B	+0.0%
_private.Init.Data.Range.Polymorphic.SInt.0.Int64.instRxcHasSize_eq	23.377 s	22.926 s	-1.9%	84.400	86.060	+2.0%	50.25 GiB	50.04 GiB	-0.4%	3.432 s	3.454 s	+0.6%	182.2 ms	183.7 ms	+0.8%	26.20 MiB	25.72 MiB	-1.8%	15.71B	15.71B	+0.0%
List.mergeSort	15.550 s	15.308 s	-1.6%	102.760	104.390	+1.6%	31.81 GiB	31.73 GiB	-0.3%	2.273 s	2.224 s	-2.2%	165.0 ms	168.9 ms	+2.3%	26.25 MiB	25.77 MiB	-1.8%	10.46B	10.46B	+0.0%
Vector.append	5.227 s	5.070 s	-3.0% 🟢	108.290	111.640	+3.1% 🟢	8.45 GiB	8.45 GiB	+0.0%	596.0 ms	590.1 ms	-1.0%	164.8 ms	163.8 ms	-0.6%	24.84 MiB	24.38 MiB	-1.8%	2.21B	2.21B	+0.0%
Nat.gcd_comm	4.440 s	4.299 s	-3.2% 🟢	93.460	96.540	+3.3% 🟢	7.92 GiB	7.56 GiB	-4.5% 🟢	471.7 ms	472.2 ms	+0.1%	177.9 ms	153.7 ms	-13.6% (1.16× faster) 🟢	24.44 MiB	23.99 MiB	-1.8%	1.56B	1.56B	+0.0%
String.append	3.259 s	3.124 s	-4.1% 🟢	108.330	112.990	+4.3% 🟢	4.92 GiB	4.94 GiB	+0.3%	348.8 ms	348.0 ms	-0.2%	148.6 ms	153.9 ms	+3.5% ⚠️	23.73 MiB	23.30 MiB	-1.8%	869.46M	869.46M	+0.0%
Int.gcd	2.717 s	2.561 s	-5.8% (1.06× faster) 🟢	84.270	89.420	+6.1% (1.06× faster) 🟢	4.80 GiB	6.08 GiB	+26.8% (1.27× larger) ⚠️	293.3 ms	287.1 ms	-2.1%	149.5 ms	138.5 ms	-7.4% (1.08× faster) 🟢	23.35 MiB	22.92 MiB	-1.8%	539.56M	539.56M	+0.0%
Nat.sub_le_of_le_add	2.500 s	2.429 s	-2.8%	76.010	78.220	+2.9%	4.92 GiB	4.69 GiB	-4.7% 🟢	283.3 ms	276.5 ms	-2.4%	149.1 ms	151.0 ms	+1.2%	23.82 MiB	23.39 MiB	-1.8%	451.92M	451.92M	+0.0%
Nat.add_comm	1.367 s	1.371 s	+0.3%	37.300	37.200	-0.3%	4.27 GiB	4.03 GiB	-5.7% (1.06× smaller) 🟢	195.7 ms	197.9 ms	+1.1%	142.5 ms	130.2 ms	-8.6% (1.09× faster) 🟢	22.27 MiB	21.85 MiB	-1.9%	48.99M	48.99M	+0.0%

13 constants · 3 with regressions · 11 with improvements (|Δ| > 3.0% on any metric).

Workflow logs

[gabriel-barrett]

!benchmark aiur-recursive fresh

[argument-ci-bot]

!benchmark — main vs c0b3b5a

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

aiur-recursive · InitStd — main from: base run @ 6e29ac9 (fresh — bencher bypassed)

proof	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%
factorial-q100-b2	5.219 s	5.063 s	-3.0% 🟢	14.56 GiB	14.58 GiB	+0.1%	11.38 MiB	10.91 MiB	-4.1% 🟢	79.3 ms	75.2 ms	-5.3% (1.06× faster) 🟢	471.7 ms	462.9 ms	-1.9%	3.88B	3.94B	+1.7%	99.1 ms	97.5 ms	-1.7%	848.62 KiB	846.23 KiB	-0.3%	7.0 ms	7.1 ms	+1.2%	373.36 MiB	363.29 MiB	-2.7%
square-q100-b1	3.796 s	3.892 s	+2.5%	9.57 GiB	9.61 GiB	+0.4%	11.31 MiB	10.84 MiB	-4.2% 🟢	78.7 ms	73.3 ms	-6.9% (1.07× faster) 🟢	424.0 ms	410.4 ms	-3.2% 🟢	3.38B	3.29B	-2.7%	88.1 ms	79.7 ms	-9.5% (1.10× faster) 🟢	782.06 KiB	779.67 KiB	-0.3%	6.5 ms	6.2 ms	-3.4% 🟢	252.83 MiB	246.45 MiB	-2.5%

2 proofs · 0 with regressions · 2 with improvements (|Δ| > 3.0% on any metric).

Workflow logs