Fast Polynomial Evaluation for Correctly Rounded Elementary Functions using the RLIBM Approach

Fast Polynomial Evaluation for Correctly Rounded Elementary Functions using the RLIBM Approach
复制标题

使用 RLIBM 方法对正确舍入的初等函数进行快速多项式计算

DOI:
10.1145/3579990.3580022
复制
发表时间:
2023
期刊:
CGO 2023: Proceedings of the 21st ACM/IEEE International Symposium on Code Generation and OptimizationFebruary 2023
影响因子:
--
通讯作者:
Nagarakatte, Santosh
Nagarakatte, Santosh
中科院分区:
--
文献类型:
--
作者:
Aanjaneya, Mridul;Nagarakatte, Santosh

文献摘要

参考文献

被引文献

相似文献

本文提出了快速多项式计算方法,使用我们的RLibm方法生成的正确舍入的初等函数。所产生的函数为具有多种表示和舍入模式的所有输入生成正确的结果。给定一个预言,RLibm方法逼近正确舍入的结果,而不是初等函数的真实的值。一个关键的观察是,在正确舍入的结果周围存在真实的值的区间,使得其中的任何真实的值舍入为正确的结果。这个区间是RLibm的多项式生成过程可用的最大自由度。随后,使用这些区间生成正确舍入的初等函数的问题可以被构造为线性规划问题。我们以前的工作RLibm的方法使用霍纳的方法多项式evaluation.This探讨多项式的评价技术,如高德纳的系数自适应程序,并行执行的操作使用埃斯特林的程序,并使用融合乘加操作的上下文中的RLibm的方法。如果我们采用RLibm方法生成的多项式并随后执行多项式求值优化,则会由于多项式求值期间的舍入误差而导致不正确的结果。因此,我们建议在RLibm的多项式生成过程中集成快速多项式求值过程。我们新的多项式求值过程结合了并行执行与融合乘加操作,优于RLibm正确舍入函数所使用的Horner方法。我们证明了32位浮点数的多项式不仅正确,而且比RLibm中的先前函数快24%
This paper proposes fast polynomial evaluation methods for correctly rounded elementary functions generated using our RLibm approach. The resulting functions produce correct results for all inputs with multiple representations and rounding modes. Given an oracle, the RLibm approach approximates the correctly rounded result rather than the real value of an elementary function. A key observation is that there is an interval of real values around the correctly rounded result such that any real value in it rounds to the correct result. This interval is the maximum freedom available to RLibm’s polynomial generation procedure. Subsequently, the problem of generating correctly rounded elementary functions using these intervals can be structured as a linear programming problem. Our prior work on the RLibm approach uses Horner’s method for polynomial evaluation.This paper explores polynomial evaluation techniques such as Knuth’s coefficient adaptation procedure, parallel execution of operations using Estrin’s procedure, and the use of fused multiply-add operations in the context of the RLibm approach. If we take the polynomial generated by the RLibm approach and subsequently perform polynomial evaluation optimizations, it results in incorrect results due to rounding errors during polynomial evaluation. Hence, we propose to integrate the fast polynomial evaluation procedure in the RLibm’s polynomial generation process. Our new polynomial evaluation procedure that combines parallel execution with fused multiply-add operations outperforms the Horner’s method used by RLibm’s correctly rounded functions. We show the resulting polynomials for 32-bit float are not only correct but also faster than prior functions in RLibm by 24%
DOI: 10.1007/978-1-4899-7983-4
发表时间: 2006
期刊: --
影响因子: --
作者:
通讯作者: --
DOI: 10.1109/tc.2010.152
发表时间: 2011
影响因子: 3.7
作者:
C. Jeannerod;Herve Knochel;Christophe Monat;G. Revy
通讯作者: G. Revy
使用融合乘法加法进行正式验证的参数缩减
DOI: --
发表时间: 2007
影响因子: 3.7
作者:
S. Boldo;M. Daumas;Ren
通讯作者: Ren
CR-LIBM:正确舍入的基本函数库
DOI: 10.1117/12.505591
发表时间: 2003
影响因子: --
作者:
Catherine Daramy;D. Defour;F. de Dinechin;J. Muller
通讯作者: J. Muller
用于改进 IA-64 上超越函数的新算法
DOI: 10.1109/arith.1999.762822
发表时间: 1999
期刊: Proceedings 14th IEEE Symposium on Computer Arithmetic (Cat. No.99CB36336)
影响因子: --
作者:
S. Story;P. T. P. Tang
通讯作者: P. T. P. Tang