Accurate Sampling with Noisy Forces from Approximate Computing

被引：6
作者：
Rengaraj, Varadarajan [1 ,2 ]
Lass, Michael [2 ,3 ]
Plessl, Christian [2 ,3 ]
Kuhne, Thomas D. [1 ,4 ]
机构：
[1] Paderborn Univ, Dept Chem, Dynam Condensed Matter, Warburger Str 100, D-33098 Paderborn, Germany
[2] Paderborn Univ, Dept Comp Sci, Warburger Str 100, D-33098 Paderborn, Germany
[3] Paderborn Univ, Paderborn Ctr Parallel Comp, Warburger Str 100, D-33098 Paderborn, Germany
[4] Paderborn Univ, Ctr Sustainable Syst Design, Warburger Str 100, D-33098 Paderborn, Germany
来源：
COMPUTATION | 2020年 / 8卷 / 02期
基金：
欧洲研究理事会;
关键词：
approximate computing; CP2K; fluctuation-dissipation theorem; FPGA; i-PI; low precision arithmetic; MOLECULAR-DYNAMICS; COMPUTATIONS; SIMULATIONS; HYBRID; MODELS; SYSTEM; JANUS;
D O I：
10.3390/computation8020039
中图分类号：
O1 [数学];
学科分类号：
0701 ; 070101 ;
摘要：
In scientific computing, the acceleration of atomistic computer simulations by means of custom hardware is finding ever-growing application. A major limitation, however, is that the high efficiency in terms of performance and low power consumption entails the massive usage of low precision computing units. Here, based on the approximate computing paradigm, we present an algorithmic method to compensate for numerical inaccuracies due to low accuracy arithmetic operations rigorously, yet still obtaining exact expectation values using a properly modified Langevin-type equation.
引用
收藏
页数：11
相关论文
共 75 条
[71]  
Strzodka R., 2006, P 14 ANN IEEE S FIEL, DOI [10.1109/FCCM.2006.57, DOI 10.1109/FCCM.2006.57]
[72]  
The Next Platform, TEARING APART GOOGLE
[73]  
Top 500, INT LAYS OUT NEW ROA
[74]   REVERSIBLE MULTIPLE TIME SCALE MOLECULAR-DYNAMICS [J].
TUCKERMAN, M ;
BERNE, BJ ;
MARTYNA, GJ .
JOURNAL OF CHEMICAL PHYSICS, 1992, 97 (03) :1990-2001
[75]   Performance potential for simulating spin models on GPU [J].
Weigel, Martin .
JOURNAL OF COMPUTATIONAL PHYSICS, 2012, 231 (08) :3064-3082
12345678