Answers are grounded in this paper and cite their sources.
Ootomo, Ozaki, and Yokota [ Int. J. High Perform. Comput. Appl., 38 (2024), pp. 297–313] have proposed a strategy to recast a floating-point matrix multiplication in terms of integer matrix products. The factors [Formula: see text] and [Formula: see text] are split into integer slices, the product of these slices is computed exactly, and [Formula: see text] is approximated by accumulating these integer products in floating-point arithmetic. This technique is particularly well suited to mixed-precision matrix multiply–accumulate units with integer support, such as the NVIDIA tensor cores or the AMD matrix cores. The number of slices allows for performance-accuracy tradeoffs: more slices yield better accuracy but require more multiplications, which in turn reduce performance. We propose an inexpensive way to estimate the minimum number of multiplications needed to achieve a prescribed level of accuracy. Our error analysis shows that the algorithm may become inaccurate (or inefficient) if rows of [Formula: see text] or columns of [Formula: see text] are badly scaled. We perform a range of numerical experiments, both in simulation and on the latest NVIDIA GPUs, that confirm the analysis and illustrate strengths and weaknesses of the algorithm. Reproducibility of computational results. This paper has been awarded the “SIAM Reproducibility Badge: Code and data available” as a recognition that the authors have followed reproducibility principles valued by SISC and the scientific computing community. Code and data that allow readers to reproduce the results in this paper are available at https://github.com/north-numerical-computing/integer-matrix-multiply-experiments and in the supplementary materials ( integer-matrix-multiply-experiments-main.zip [97.8KB]). [Formula: see text]
No takes yet. Share an insight, caveat, or question.
Abdelfattah et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: