TY - GEN
T1 - From Molecular Dynamics to Oceanography - Ookami Graduate Students Porting and Tuning Science Codes for A64FX
AU - Kaushik, Kedarsh
AU - Wang, Yuzhang
AU - Ma, Youwei
AU - Carlson, David
AU - Curtis, Tony
AU - Harrison, Robert
AU - Siegmann, Eva
N1 - Publisher Copyright:
© 2023 Owner/Author.
PY - 2023/7/23
Y1 - 2023/7/23
N2 - To build a cohort of computationally educated young researchers and to deepen community expertise, the Ookami project is actively integrating graduate students into the project team. They learn how to get started on a novel architecture and what to take care of when porting applications. This experience enables them to do production research on A64FX as well as grow their general knowledge in HPC, benefiting not just them but also the codes they are working on. Embedding them within the project team greatly facilitates the transfer of skills, builds a common vocabulary and culture, provides experience in large project execution, and builds confidence and a can-do attitude. In this paper, we report the student's preliminary efforts in porting several science applications to Fujitsu A64FX. The Arm-based Fujitsu A64FX processor developed by Fujitsu and RIKEN is used in Fugaku, which until June 2022 has been the fastest machine worldwide for two years. Its main features of SVE, HBM, and being power efficient makes it unique in the world of HPC. The studied applications (SIESTA, MOM6, Amber, ROMS) are from various disciplines and give a good overview of the A64FX porting process.
AB - To build a cohort of computationally educated young researchers and to deepen community expertise, the Ookami project is actively integrating graduate students into the project team. They learn how to get started on a novel architecture and what to take care of when porting applications. This experience enables them to do production research on A64FX as well as grow their general knowledge in HPC, benefiting not just them but also the codes they are working on. Embedding them within the project team greatly facilitates the transfer of skills, builds a common vocabulary and culture, provides experience in large project execution, and builds confidence and a can-do attitude. In this paper, we report the student's preliminary efforts in porting several science applications to Fujitsu A64FX. The Arm-based Fujitsu A64FX processor developed by Fujitsu and RIKEN is used in Fugaku, which until June 2022 has been the fastest machine worldwide for two years. Its main features of SVE, HBM, and being power efficient makes it unique in the world of HPC. The studied applications (SIESTA, MOM6, Amber, ROMS) are from various disciplines and give a good overview of the A64FX porting process.
KW - A64FX
KW - Amber
KW - high performance computing
KW - MOM6
KW - porting
KW - ROMS
KW - SIESTA
UR - https://www.scopus.com/pages/publications/85176223516
U2 - 10.1145/3569951.3593608
DO - 10.1145/3569951.3593608
M3 - Conference contribution
AN - SCOPUS:85176223516
T3 - PEARC 2023 - Computing for the common good: Practice and Experience in Advanced Research Computing
SP - 18
EP - 25
BT - PEARC 2023 - Computing for the common good
PB - Association for Computing Machinery, Inc
T2 - 2023 Practice and Experience in Advanced Research Computing, PEARC 2023
Y2 - 23 July 2023 through 27 July 2023
ER -