This research investigates using a mixed-precision iterative refinement method using posit numbers instead of the standard IEEE floating-point format. The method is applied to solve a general linear system represented by the equation Ax=b, where A is a large sparse matrix. Various scaling techniques, such as row and column equilibration, map the matrix entries to higher-density regions of machine numbers before performing the O(n3) factorization operation. Low-precision LU factorization followed by forward/backward substitution provides an initial estimate. The results demonstrate that a 16-bit posit configuration combined with equilibration produces accuracy comparable to IEEE half-precision (fp16), indicating a potential for achieving a balance between efficiency and accuracy.
@article{arxiv.2408.13400,
title = {Iterative Refinement with Low-Precision Posits},
author = {James Quinlan and E. Theodore L. Omtzigt},
journal= {arXiv preprint arXiv:2408.13400},
year = {2024}
}