Fast Mixed-Precision Real Evaluation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yadrov, Artem, Panchekha, Pavel
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912333480067072
author Yadrov, Artem
Panchekha, Pavel
author_facet Yadrov, Artem
Panchekha, Pavel
contents Evaluating real-valued expressions to high precision is a key building block in computational mathematics, physics, and numerics. A typical implementation evaluates the whole expression in a uniform precision, doubling that precision until a sufficiently-accurate result is achieved. This is wasteful: usually only a few operations really need to be performed at high precision, and the bulk of the expression could be computed much faster. However, such non-uniform precision assignments have, to date, been impractical to compute. We propose a fast new algorithm for deriving such precision assignments. The algorithm leverages results computed at lower precisions to analytically determine a mixed-precision assignment that will result in a sufficiently-accurate result. Our implementation, Reval, achieves an average speed-up of 1.72x compared to the state-of-the-art Sollya tool, with the speed-up increasing to 5.21x on the most difficult input points. An examination of the precisions used with and without precision tuning shows that the speed-up results from assigning lower precisions for the majority of operations, though additional optimizations enabled by the non-uniform precision assignments also play a role.
format Preprint
id arxiv_https___arxiv_org_abs_2504_11708
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Fast Mixed-Precision Real Evaluation
Yadrov, Artem
Panchekha, Pavel
Numerical Analysis
Mathematical Software
Evaluating real-valued expressions to high precision is a key building block in computational mathematics, physics, and numerics. A typical implementation evaluates the whole expression in a uniform precision, doubling that precision until a sufficiently-accurate result is achieved. This is wasteful: usually only a few operations really need to be performed at high precision, and the bulk of the expression could be computed much faster. However, such non-uniform precision assignments have, to date, been impractical to compute. We propose a fast new algorithm for deriving such precision assignments. The algorithm leverages results computed at lower precisions to analytically determine a mixed-precision assignment that will result in a sufficiently-accurate result. Our implementation, Reval, achieves an average speed-up of 1.72x compared to the state-of-the-art Sollya tool, with the speed-up increasing to 5.21x on the most difficult input points. An examination of the precisions used with and without precision tuning shows that the speed-up results from assigning lower precisions for the majority of operations, though additional optimizations enabled by the non-uniform precision assignments also play a role.
title Fast Mixed-Precision Real Evaluation
topic Numerical Analysis
Mathematical Software
url https://arxiv.org/abs/2504.11708