Sorting networks on 13 inputs need at least 44 comparators

Let S(N) be the least number of comparators in a sorting network on N inputs (the Bose–Nelson problem). The values are known for N ≤ 12, and the best proved bounds for 13 inputs have been 43 ≤ S(13) ≤ 45. We prove S(13) ≥ 44. Deleting the comparators met by the largest and the second largest input leaves a sorting network on N − 2 inputs, so S(N) ≥ S(N − 2) + P(2, N), where P(2, N) is the least possible worst-case number of comparators met by these two values. In 1972 Van Voorhis stated P(2, N) ≥ ⌈log₂ F(N)⌉ for an explicit function F, but a 2026 audit showed that his proof is invalid. We determine P(2, N) by computer instead. We show that the number of comparators met by the two values in each run is determined by the binary tree formed by the paths of the maximum, the comparators that these paths pass, and the depths at which the second maximum enters a forest of further comparators. This reduces the inequality P(2, N) ≤ L to a finite search. For N = 13 and L = 8 only 25 cases remain, and none of them can be completed, so P(2, 13) = 9, and Harder's S(11) = 35 gives S(13) ≥ 44. Independent checks confirm P(2, 13) = 9: a second implementation with its own lemmas, an exhaustive search over networks that does not use our structural reduction, and an integer program for each case. The same searches give P(2, N) = ⌈log₂ F(N)⌉ for 3 ≤ N ≤ 27, which proves the lower bounds on S(N) for 13 ≤ N ≤ 27 listed in Dobbelaere's public table. Programs, logs and independent checks: doi:10.5281/zenodo.23453295 (repository https://github.com/tejasvi/sortnet13-two-value). Mathematics Subject Classification (2020): 68P10 (primary); 68Q06, 05C05, 68V05 (secondary).

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-10-18
DOI
https://doi.org/10.5281/zenodo.23453292
Primary Topic
Complexity and Algorithms in Graphs
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
preprint

Sorting networks on 13 inputs need at least 44 comparators

Tejasvi Singh Tomar
Zenodo (CERN European Organization for Nuclear Research)
Complexity and Algorithms in Graphs
preprint

Sorting networks on 13 inputs need at least 44 comparators

Tejasvi Singh Tomar
preprint en

Abstract

Let S(N) be the least number of comparators in a sorting network on N inputs (the Bose–Nelson problem). The values are known for N ≤ 12, and the best proved bounds for 13 inputs have been 43 ≤ S(13) ≤ 45. We prove S(13) ≥ 44. Deleting the comparators met by the largest and the second largest input leaves a sorting network on N − 2 inputs, so S(N) ≥ S(N − 2) + P(2, N), where P(2, N) is the least possible worst-case number of comparators met by these two values. In 1972 Van Voorhis stated P(2, N) ≥ ⌈log₂ F(N)⌉ for an explicit function F, but a 2026 audit showed that his proof is invalid. We determine P(2, N) by computer instead. We show that the number of comparators met by the two values in each run is determined by the binary tree formed by the paths of the maximum, the comparators that these paths pass, and the depths at which the second maximum enters a forest of further comparators. This reduces the inequality P(2, N) ≤ L to a finite search. For N = 13 and L = 8 only 25 cases remain, and none of them can be completed, so P(2, 13) = 9, and Harder's S(11) = 35 gives S(13) ≥ 44. Independent checks confirm P(2, 13) = 9: a second implementation with its own lemmas, an exhaustive search over networks that does not use our structural reduction, and an integer program for each case. The same searches give P(2, N) = ⌈log₂ F(N)⌉ for 3 ≤ N ≤ 27, which proves the lower bounds on S(N) for 13 ≤ N ≤ 27 listed in Dobbelaere's public table. Programs, logs and independent checks: doi:10.5281/zenodo.23453295 (repository https://github.com/tejasvi/sortnet13-two-value). Mathematics Subject Classification (2020): 68P10 (primary); 68Q06, 05C05, 68V05 (secondary).

Zenodo (CERN European Organization for Nuclear Research)
Complexity and Algorithms in Graphs
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.