Answers are grounded in this paper and cite their sources.
Every entry of the modulo-30 wheel is expressed through the offsets ρ ∈ {0, 4, 6, 10, 12, 16, 22, 24}of its eight vectors, and every product of two entries as N = s2 − d2, where s = 7 + α + 15kand d = β + 15j are linear in the crossing diagonal k = m + n and the diagonal j = n − m ofthe two factors (formulas of the first author). On these coordinates a base table of 36 grids isconstructed, one for each family F(a, b), in which every product of two wheel entries occupiesexactly one cell that yields p, q, s, d and N simultaneously. The parametrisation, the familiesof each vector, the position law, an ordered form with constant second difference 1800, thelayers of the position, the diagonal element associated with each N, and a formulation ofthe complete system in the single unknown k are established. For the hard semiprimes thediagonals are shown to appear in periods of increasing length, Bt+1 = Bt + 2C∆L2; thehalf-difference satisfies d ≤ 3.840√n (verified to 1014), and an exact generator of all hardsemiprimes in increasing order is obtained. For a given N the factorisation is the uniquelattice point of the hyperbola s2 − d2 = N in its grid. It is located by parallel streams, byrotating residue matrices, and by a baby-step giant-step search in k based on the congruence4s ≡ 2N+1 (mod N); a single search shared by all families lowers the worst-case cost by afactor between 2 and 2.83, and 57 hard semiprimes of 20 to 40 digits are factored. Threewindow methods are introduced: an adaptive wheel table whose length, a multiple of aprimorial, is placed near the expected smaller factor (70 digits in 34 seconds); a dynamicsieve on a growing wheel 30, 210, 2310, . . . that retains the eight levels of the wheel 30 andclassifies a window completely without a probabilistic test, with up to 1.73 times fewerfactors than the wheel 30; and the base table in one variable, E = X +Dk + 30n, coloured bydirectional multiplication, whose bit-level implementation is 3.7 times faster. A window of30,000 integers at 9 · 1029 is classified completely: its 425 primes and its 316 semiprimes withboth factors above 1010 are determined deterministically, in agreement with the expecteddensities in every period. The method determines an interval rather than a single numberand runs in independent periods on ordinary processors. A hard semiprime of 24 digits isfactored on the same table in 504 seconds.
No takes yet. Share an insight, caveat, or question.
Msliki et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: