Finite sieve coverage of the Richstein range in blocks of one million
OpenRichstein2001.segmented_sieve_coveragecomputational-number-theorygoldbachnumber-theory
Let , , , , and let be the finite survivor set defined in GoldbachSieve: numbers at least 2 in with no prime divisor at most other than themselves.
For every integer , put and . The finite coverage assertion is
This is the outstanding computational obligation in a sieve-based formalization of Richstein's verification. The original abstract reports the range and maximal smaller prime 5569. The block width and certificate interface are formalization choices. No certificate data or Lean verification of all these blocks is supplied with this statement. The last block includes the endpoint ; the first block includes .
Preamble
import Definitions.Def_GoldbachSieve import Mathlib.Algebra.Ring.Parity
Formal statement
namespace Richstein2001
/-- Finite segmented sieve coverage obligation; the computational data remain open. -/
theorem segmented_sieve_coverage (b : ℕ) (hb : b < 400000001) :
((Finset.Icc (max 4 (b * 1000000))
(min (4 * 10 ^ 14) ((b + 1) * 1000000 - 1))).filter (fun n => Even n)) ⊆
GoldbachSieve.pairSums 5569 (b * 1000000 - 5569)
(min (4 * 10 ^ 14) ((b + 1) * 1000000 - 1)) 20000000 := by sorry
end Richstein2001
Source
J. Richstein, Verifying the Goldbach conjecture up to 4·10^14, Math. Comp. 70 (2001), 1745–1749; abstract p. 1745 reports segmented sieving and maximal smaller prime 5569. https://doi.org/10.1090/S0025-5718-00-01290-4 . The finite-set interface and width 10^6 are this formalization's choices, not a transcription of the original program or recovered certificates.