arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2012.00706math.NAcs.NAphysics.comp-ph
收藏

模型降阶的混合精度矩阵插值分解

Mixed precision matrix interpolative decompositions for model reduction

Alec Michael Dunton, Alyson Fox

摘要:

由于数据容量和带宽的不断增长以及GPU的进步,人们对混合精度算法重新产生了兴趣,这使得低精度算法能够显着加速。鉴于此,我们提出了一种混合精度算法,以产生一个双精度精确矩阵插值分解近似在一组给定的标准。尽管低精度算法遭受更快的舍入误差累积,但是对于许多数据丰富的应用,我们仍然获得可行的近似精度,因为使用低精度算法引起的误差由低秩近似固有的误差主导。然后,我们进行了几个模拟数值试验,以证明算法的有效性和相应的误差估计。最后,我们提出了我们的算法的应用程序中的一个问题,在模型减少颗粒载货湍流。

英文摘要:

Renewed interest in mixed-precision algorithms has emerged due to growing data capacity and bandwidth concerns, as well as the advancement of GPUs, which enable significant speedup for low precision arithmetic. In light of this, we propose a mixed-precision algorithm to generate a double-precision accurate matrix interpolative decomposition approximation under a given set of criteria. Though low precision arithmetic suffers from quicker accumulation of round-off error, for many data-rich applications we nevertheless attain viable approximation accuracy, as the error incurred using low precision arithmetic is dominated by the error inherent to low-rank approximation. We then conduct several simulated numerical tests to demonstrate the efficacy of the algorithms and the corresponding error estimates. Finally, we present the application of our algorithms to a problem in model reduction for particle-laden turbulent flow.

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。

↑