arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2107.01604math.NAcs.NA
收藏

浮点求和算法的确定性和概率性误差界

Deterministic and Probabilistic Error Bounds for Floating Point Summation Algorithms

Eric Hallman, Ilse C. F. Ipsen

摘要:

我们分析了前向误差的浮点求和的实数,从算法,不需要诉诸更高的精度或更好的硬件。我们推导出翔实的显式表达式,和新的确定性和概率界的错误三类算法:一般求和,移位一般求和,补偿(顺序)求和。我们的概率界一般和移位一般求和举行的所有订单。对于补偿求和,我们还提出了确定性和概率的一阶和二阶界限,与一阶界限,不同于现有的。数值实验表明,边界是信息和三个算法类,补偿求和一般是最准确的方法。

英文摘要:

We analyse the forward error in the floating point summation of real numbers, from algorithms that do not require recourse to higher precision or better hardware. We derive informative explicit expressions, and new deterministic and probabilistic bounds for errors in three classes of algorithms: general summation,shifted general summation, and compensated (sequential) summation. Our probabilistic bounds for general and shifted general summation hold to all orders. For compensated summation, we also present deterministic and probabilistic first and second order bounds, with a first order bound that differs from existing ones. Numerical experiments illustrate that the bounds are informative and that among the three algorithm classes, compensated summation is generally the most accurate method.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑