Paper
19 October 2022 Hot spot locating and optimization of calculation errors in arithmetic expressions
Jinchen Xu, Yuanyuan Xia, Jiangwei Hao, Hao Cao
Author Affiliations +
Proceedings Volume 12294, 7th International Symposium on Advances in Electrical, Electronics, and Computer Engineering; 122945M (2022) https://doi.org/10.1117/12.2639915
Event: 7th International Symposium on Advances in Electrical, Electronics and Computer Engineering (ISAEECE 2022), 2022, Xishuangbanna, China
Abstract
Error analysis is an important support for improving the precision of floating-point calculations. However, the floatingpoint error distribution does not have dominant characteristics, making it difficult to optimize floating-point calculation precision. This paper proposes the LFEE (Locating hot spots of floating-point expression errors) algorithm. The algorithm analyzes the error detection result of the expression to obtain the error hot spot, i.e., the step with the maximum error in the expression calculation sequence, followed by preforming precision optimization scheme for the error hot spot. The experimental results show that the LFEE algorithm can accurately locate the error hot spots and improve the accuracy of the function by optimizing the error hot spots.
© (2022) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Jinchen Xu, Yuanyuan Xia, Jiangwei Hao, and Hao Cao "Hot spot locating and optimization of calculation errors in arithmetic expressions", Proc. SPIE 12294, 7th International Symposium on Advances in Electrical, Electronics, and Computer Engineering, 122945M (19 October 2022); https://doi.org/10.1117/12.2639915
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Error analysis

Optimization (mathematics)

Numerical stability

Back to Top