Apps · PROJECT NOTES
AlgoTune
A benchmark and execution harness for optimizing numerical and algorithmic programs while preserving valid outputs, measured against reference runtimes.
What can you do with it?
Measure correctness-preserving solver speedups on generated held-out inputs, using task-specific generators, validators and reference solvers.
What should you know before using it?
Timing depends on hardware, library versions, calibration and parallel contention. The 154 task families each generate many inputs; input counts are not additional tasks. Original aggregate speedup, AlgoTune-Pass resolve rate and Harbor rewards are different measures. Harbor documents a 1.0 floor for invalid or slower solutions, so that reward alone does not prove correctness.
What is verified here?
Runtime: Python >=3.10 and numerical dependencies; documented launchers support standalone, SLURM and optional AWS paths. AlgoTuner requires model configuration. Preserve train/test input seeds, problem-size calibration, CPU/thread limits and correctness traces. No solver, timing benchmark or cloud resource was run.
License note: Repository benchmark/harness code is MIT. Numerical libraries and other dependencies retain their own licenses. Upstream license ↗
This entry was reviewed against original project material. We have not independently run the software or reproduced its results. AI-development claims are author-reported; check upstream for current versions, setup requirements, and licenses.
- Resource type
- Benchmark suite
- Provenance
- Development provenance not established
- Reviewed
- Platforms
- Consult upstream documentation