Published March 9, 2026
| Version v1
Dataset
Open
Tide Max Flow Algorithm — Benchmark Data and Graph Pool
Authors/Creators
Description
Benchmark data and graph pool for the paper Tide: A Global Push-Pull-Relabel Algorithm for Maximum Flow.
Contains:
- 79 DIMACS max-flow graph instances across 8 families (layered, grid, chains, bipartite, random, Washington, RFIM, vision) totaling ~4.9 GB uncompressed
- Raw benchmark timing data: 9 solver configurations × 5 runs each (45 CSV files + collated JSON)
- Benchmark driver scripts and solver wrappers for reproducibility
See README.md for full details on graph families, solvers, and methodology.
Notes
Platform: Apple M4 Pro, 48 GB RAM, macOS 26.3. Rust 1.93.0, Python 3.14.2. All max-flow values cross-validated across all 9 solver configurations.
Files
README.md
Files
(755.5 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:b1d1f45f400f6c156dcd597d6c5abf5a
|
4.6 kB | Preview Download |
|
md5:081c24d215bdeb103d2a0d7c094987bb
|
48.6 kB | Download |
|
md5:8020e4b22c62ab80966612605882358d
|
755.4 MB | Download |
Additional details
Related works
- Is supplemented by
- Software: https://github.com/tpapak/tide-maxflow (URL)
- Software: https://github.com/tpapak/algraph (URL)
- Software: https://crates.io/crates/tide-maxflow (URL)
- Software: https://hackage.haskell.org/package/algraph (URL)