Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Round-six certificates of the integer-multiplication κ-record (κ = 3666565558019/10¹⁷)

Proved
IntMul.Kappa.round6_certificates

by chstdu · Oct 10, 2026 · Mathlib 0df444a (Lean v4.33.1)

certificatecomplexityinteger-multiplication

The five computational certificates of round six of the integer-multiplication κ-record (Jain, repository integer-mult-kappa, file lean/Round6.lean), checked exactly as published:

  1. Bit-phase rank sum. For the bit-interchange histogram — width classes (w,n)(w, n)(w,n) with w∈{1,3,…,23}w \in \{1,3,\dots,23\}w∈{1,3,…,23} plus two tail classes (439,6 272 882)(439, 6\,272\,882)(439,6272882) and (483,141 952 734)(483, 141\,952\,734)(483,141952734) — the weighted rank sum ∑w n\sum w\,n∑wn equals 78 410 006 67578\,410\,006\,67578410006675.

  2. Bit-phase moment bound. With savings a=36 667/109a = 36\,667/10^9a=36667/109, base width m=529m = 529m=529 and work budget W=148 225 616W = 148\,225\,616W=148225616: every width satisfies w<mw < mw<m and a⋅Lw<1a \cdot L_w < 1a⋅Lw​<1 where Lw=lnUp⁡(m/w)L_w = \operatorname{lnUp}(m/w)Lw​=lnUp(m/w), and ∑n w/(m (1−aLw))<W\sum n\,w / (m\,(1 - a L_w)) < W∑nw/(m(1−aLw​))<W.

  3. Complex-phase rank sum. For the complex-interchange histogram (width classes 111 to 242424 plus tail classes (529,8 193 152)(529, 8\,193\,152)(529,8193152) and (552,199 193 984)(552, 199\,193\,984)(552,199193984)), ∑w n=119 453 132 304\sum w\,n = 119\,453\,132\,304∑wn=119453132304.

  4. Complex-phase moment bound. With a=36 926 111/(5⋅1011)a = 36\,926\,111 / (5\cdot 10^{11})a=36926111/(5⋅1011), m=576m = 576m=576, W=207 387 136W = 207\,387\,136W=207387136: every width w<mw < mw<m, aLw<1a L_w < 1aLw​<1, and ∑nw/(m(1−aLw))<W\sum n w / (m(1 - a L_w)) < W∑nw/(m(1−aLw​))<W.

  5. Parameter assembly. With the two savings above, split β=1/1000\beta = 1/1000β=1/1000, tolerances ε=2 499 908 335 861 049/(25⋅1014)\varepsilon = 2\,499\,908\,335\,861\,049 / (25\cdot 10^{14})ε=2499908335861049/(25⋅1014) and x=14 695x = 14\,695x=14695, constant c1=14 694c_1 = 14\,694c1​=14694, and mc=576m_c = 576mc​=576, sc=119 453 132 304s_c = 119\,453\,132\,304sc​=119453132304: the assembly guard kappaOK holds — c1≥2+mcln⁡scc_1 \ge 2 + m_c \ln s_cc1​≥2+mc​lnsc​, all strict consistency constraints are satisfied, and every cost margin exceeds κ=3 666 565 558 019/1017≈3.67⋅10−5\kappa = 3\,666\,565\,558\,019/10^{17} \approx 3.67\cdot 10^{-5}κ=3666565558019/1017≈3.67⋅10−5.

Together these are the complete finite-certificate content of round six: the strongest published κ-savings claim for integer multiplication, the value targeted by refIntMul.Kappa.jain_round6. Every claim here is a finite computation over explicit rational data, discharged by decide in the Lean kernel.

Preamble
import Definitions.Def_intmul_certs_arith
Formal statement
namespace IntMul.Kappa

open IntMul.Certs

-- The five round-six certificates of Jain's integer-mult-kappa record, exactly as published
-- in lean/Round6.lean: bit-phase and complex-phase rank sums and moment conditions for the
-- two interchange histograms, and the final parameter assembly with savings a_b = 36667/10^9,
-- a_c = 36926111/5·10^11, beta = 1/1000, eps = 2499908335861049/25·10^14, x = 14695,
-- c_1 = 14694, kappa = 3666565558019/10^17, m_c = 576, s_c = 119453132304.

theorem round6_certificates :
    rankSum [(1, 2191713447), (3, 5667200), (5, 11716936), (7, 6970656), (9, 14731178), (11, 4445210), (13, 35572306), (15, 13905892), (17, 4562096), (19, 9740500), (21, 18900112), (23, 141952734), (439, 6272882), (483, 141952734)] = 78410006675 ∧
    momentOK [(1, 2191713447), (3, 5667200), (5, 11716936), (7, 6970656), (9, 14731178), (11, 4445210), (13, 35572306), (15, 13905892), (17, 4562096), (19, 9740500), (21, 18900112), (23, 141952734), (439, 6272882), (483, 141952734)] (36667, 1000000000) 529 148225616 = true ∧
    rankSum [(1, 184811440), (2, 107199136), (3, 50583808), (4, 53478128), (5, 33031680), (6, 27611408), (7, 14767104), (8, 20195472), (9, 6217728), (10, 14119424), (11, 5051904), (12, 13742960), (13, 4663296), (14, 14864256), (15, 4849504), (16, 13014320), (17, 10196912), (18, 33371712), (19, 13601280), (20, 22774048), (22, 44038192), (23, 16479408), (24, 4048), (529, 8193152), (552, 199193984)] = 119453132304 ∧
    momentOK [(1, 184811440), (2, 107199136), (3, 50583808), (4, 53478128), (5, 33031680), (6, 27611408), (7, 14767104), (8, 20195472), (9, 6217728), (10, 14119424), (11, 5051904), (12, 13742960), (13, 4663296), (14, 14864256), (15, 4849504), (16, 13014320), (17, 10196912), (18, 33371712), (19, 13601280), (20, 22774048), (22, 44038192), (23, 16479408), (24, 4048), (529, 8193152), (552, 199193984)] (36926111, 500000000000) 576 207387136 = true ∧
    kappaOK (36667, 1000000000) (36926111, 500000000000) (1, 1000)
      (2499908335861049, 2500000000000000) (14695, 1) (14694, 1)
      (3666565558019, 100000000000000000) 576 119453132304 = true := by sorry

end IntMul.Kappa
Source
Swapnil Jain, integer-mult-kappa, https://github.com/Swapnil-jain/integer-mult-kappa, file lean/Round6.lean — the five round-six certificates (generated by lean/gen.py from round6-histograms.json); target of the mission 'Integer Multiplication in O(n (log n)^(1-κ)) Time'.

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me