Skip to content

north-numerical-computing/integer-matrix-multiply-experiments

Repository files navigation

MATLAB experiments with integer-bassed matrix multiply

This repository contains the data plotted in [Sec. 4, 1] and the MATLAB code used to generate it.

The scripts rely on the gemmi library for simulating integer-based GEMM.

The errors in data were produced with MATLAB version R2024a Update 2 and the gemmi version under the SHA ddc562ea9404b6ced325ef66a6a804ef383b73e5. Run the script run_tests.m to regenerate the data files in data or run the shell script plot.sh to regenerate the data files and plot the figures in diagrams.

References

[1] A. Abdelfattah, J. Dongarra, M. Fasi, M. Mikaitis, and F. Tisseur Error Analysis of Floating-Point Matrix Multiplication Computed via Low-Precision Integer Arithmetic. arXiv:2506.11277 [math.NA]. June, 2025.

License

This software is distributed under the terms of the 2-clause BSD software license (see LICENSE).

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published