TY - GEN
T1 - A Large Data Visualization Framework for SPARC64 fx HPC Systems-Case Study
T2 - 8th IEEE Symposium on Large Data Analysis and Visualization, LDAV 2018
AU - Nonaka, Jorji
AU - Ono, Kenji
AU - Sakamoto, Naohisa
AU - Hayashi, Kengo
AU - Matsuda, Motohiko
AU - Shoji, Fumiyoshi
AU - Oku, Kentaro
AU - Fujita, Masahiro
AU - Hatta, Kazuma
PY - 2018/10
Y1 - 2018/10
N2 - Leading-edge supercomputer systems have been designed to achieve the highest computational performance possible for running a wide variety of large-scale simulations, and the pre-and post-processing are usually not considered in the main design feature. Although supercomputer systems may have peculiar CPU architecture, the auxiliary computational systems tend to use commodity based hardware and software in the form of servers and clusters. In the case of the K computer operational environment, at RIKEN R-CCS, the supercomputer itself is based on SPARC64 fx CPU architecture, and pre-and post-processing servers are based on traditional x86 CPU architecture. In this poster we present a large data visualization environment developed for this peculiar HPC operational environment, presenting some of the efforts made to meet the large data visualization needs. It is publicly known that the next-generation leading-edge Japanese supercomputer will abandon this CPU architecture in favor of another architecture, but we expect that some of the knowledge obtained in this development will also be useful for this future coming supercomputer system.
AB - Leading-edge supercomputer systems have been designed to achieve the highest computational performance possible for running a wide variety of large-scale simulations, and the pre-and post-processing are usually not considered in the main design feature. Although supercomputer systems may have peculiar CPU architecture, the auxiliary computational systems tend to use commodity based hardware and software in the form of servers and clusters. In the case of the K computer operational environment, at RIKEN R-CCS, the supercomputer itself is based on SPARC64 fx CPU architecture, and pre-and post-processing servers are based on traditional x86 CPU architecture. In this poster we present a large data visualization environment developed for this peculiar HPC operational environment, presenting some of the efforts made to meet the large data visualization needs. It is publicly known that the next-generation leading-edge Japanese supercomputer will abandon this CPU architecture in favor of another architecture, but we expect that some of the knowledge obtained in this development will also be useful for this future coming supercomputer system.
UR - http://www.scopus.com/inward/record.url?scp=85068820052&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85068820052&partnerID=8YFLogxK
U2 - 10.1109/LDAV.2018.8739214
DO - 10.1109/LDAV.2018.8739214
M3 - Conference contribution
T3 - 2018 IEEE 8th Symposium on Large Data Analysis and Visualization, LDAV 2018
SP - 108
EP - 109
BT - 2018 IEEE 8th Symposium on Large Data Analysis and Visualization, LDAV 2018
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 21 October 2018
ER -