TY - GEN
T1 - Performance of the Supercomputer Fugaku for Breadth-First Search in Graph500 Benchmark
AU - Nakao, Masahiro
AU - Ueno, Koji
AU - Fujisawa, Katsuki
AU - Kodama, Yuetsu
AU - Sato, Mitsuhisa
N1 - Funding Information:
Acknowledgments. We would like to express our sincere thanks to Fujitsu engineers of the supercomputer Fugaku for helping us execute the benchmark. We are also grateful to Dr. Yutaka Ishikawa, the project leader of the Flagship 2020 Project. This work is partially funded by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) program for the Development and Improvement for the Next Generation Ultra-High-Speed Computer System, under its Subsidies for Operating the Specific Advanced Large Research Facilities. This work is also partially funded by RIKEN Incentive Research Projects.
Publisher Copyright:
© 2021, Springer Nature Switzerland AG.
PY - 2021
Y1 - 2021
N2 - In this paper, we present the performance of the supercomputer Fugaku for breadth-first search (BFS) problem in the Graph500 benchmark, which is known as a ranking benchmark used to evaluate large-scale graph processing performance on supercomputer systems. Fugaku is a huge-scale Japanese exascale supercomputer that consists of 158,976 nodes connected by the Tofu interconnect D (TofuD). We have developed a BFS implementation that can extract the performance of Fugaku. We also optimize the number of processes per node, one-to-one communication, performance power ratio, and process mapping in the six-dimensional mesh/torus topology of TofuD. We evaluate the BFS performance for a large-scale graph consisting of about 2.2 trillion vertices and 35.2 trillion edges using the whole Fugaku system, and achieve 102,956 giga-traversed edges per second (GTEPS), resulting in the first position of Graph500 BFS ranking in November 2020. This performance is 3.3 times higher than that of Fugaku’s previous system, the K computer.
AB - In this paper, we present the performance of the supercomputer Fugaku for breadth-first search (BFS) problem in the Graph500 benchmark, which is known as a ranking benchmark used to evaluate large-scale graph processing performance on supercomputer systems. Fugaku is a huge-scale Japanese exascale supercomputer that consists of 158,976 nodes connected by the Tofu interconnect D (TofuD). We have developed a BFS implementation that can extract the performance of Fugaku. We also optimize the number of processes per node, one-to-one communication, performance power ratio, and process mapping in the six-dimensional mesh/torus topology of TofuD. We evaluate the BFS performance for a large-scale graph consisting of about 2.2 trillion vertices and 35.2 trillion edges using the whole Fugaku system, and achieve 102,956 giga-traversed edges per second (GTEPS), resulting in the first position of Graph500 BFS ranking in November 2020. This performance is 3.3 times higher than that of Fugaku’s previous system, the K computer.
UR - http://www.scopus.com/inward/record.url?scp=85111430735&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85111430735&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-78713-4_20
DO - 10.1007/978-3-030-78713-4_20
M3 - Conference contribution
AN - SCOPUS:85111430735
SN - 9783030787127
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 372
EP - 390
BT - High Performance Computing - 36th International Conference, ISC High Performance 2021, Proceedings
A2 - Chamberlain, Bradford L.
A2 - Chamberlain, Bradford L.
A2 - Varbanescu, Ana-Lucia
A2 - Ltaief, Hatem
A2 - Luszczek, Piotr
PB - Springer Science and Business Media Deutschland GmbH
T2 - 36th International Conference on High Performance Computing, ISC High Performance 2021
Y2 - 24 June 2021 through 2 July 2021
ER -