TY - GEN
T1 - Pushing the limit of molecular dynamics with ab initio accuracy to 100 million atoms with machine learning
AU - Jia, Weile
AU - Wang, Han
AU - Chen, Mohan
AU - Lu, Denghui
AU - Lin, Lin
AU - Car, Roberto
AU - Weinan, E.
AU - Zhang, Linfeng
N1 - Publisher Copyright:
© 2020 IEEE.
PY - 2020/11
Y1 - 2020/11
N2 - For 35 years, ab initio molecular dynamics (AIMD) has been the method of choice for modeling complex atomistic phenomena from first principles. However, most AIMD applications are limited by computational cost to systems with thousands of atoms at most. We report that a machine learningbased simulation protocol (Deep Potential Molecular Dynamics), while retaining ab initio accuracy, can simulate more than 1 nanosecond-long trajectory of over 100 million atoms per day, using a highly optimized code (GPU DeePMD-kit) on the Summit supercomputer. Our code can efficiently scale up to the entire Summit supercomputer, attaining 91 PFLOPS in double precision (45.5% of the peak) and 162/275 PFLOPS in mixed-single/half precision. The great accomplishment of this work is that it opens the door to simulating unprecedented size and time scales with ab initio accuracy. It also poses new challenges to the next-generation supercomputer for a better integration of machine learning and physical modeling.
AB - For 35 years, ab initio molecular dynamics (AIMD) has been the method of choice for modeling complex atomistic phenomena from first principles. However, most AIMD applications are limited by computational cost to systems with thousands of atoms at most. We report that a machine learningbased simulation protocol (Deep Potential Molecular Dynamics), while retaining ab initio accuracy, can simulate more than 1 nanosecond-long trajectory of over 100 million atoms per day, using a highly optimized code (GPU DeePMD-kit) on the Summit supercomputer. Our code can efficiently scale up to the entire Summit supercomputer, attaining 91 PFLOPS in double precision (45.5% of the peak) and 162/275 PFLOPS in mixed-single/half precision. The great accomplishment of this work is that it opens the door to simulating unprecedented size and time scales with ab initio accuracy. It also poses new challenges to the next-generation supercomputer for a better integration of machine learning and physical modeling.
KW - Deep potential molecular dynamics
KW - GPU
KW - Summit
KW - ab initio molecular dynamics
KW - heterogeneous architecture
KW - machine learning
UR - http://www.scopus.com/inward/record.url?scp=85098637644&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85098637644&partnerID=8YFLogxK
U2 - 10.1109/SC41405.2020.00009
DO - 10.1109/SC41405.2020.00009
M3 - Conference contribution
AN - SCOPUS:85098637644
T3 - International Conference for High Performance Computing, Networking, Storage and Analysis, SC
BT - Proceedings of SC 2020
PB - IEEE Computer Society
T2 - 2020 International Conference for High Performance Computing, Networking, Storage and Analysis, SC 2020
Y2 - 9 November 2020 through 19 November 2020
ER -