TY - GEN
T1 - Performance Analysis of CPU and DRAM Power Constrained Systems with Magnetohydrodynamic Simulation Code
AU - Fukazawa, Keiichiro
AU - Ueda, Masatsugu
AU - Inadomi, Yuichi
AU - Aoyagi, Mutsumi
AU - Umeda, Takayuki
AU - Inoue, Koji
N1 - Funding Information:
ACKNOWLEDGMENT This work was supported by Grant-in-Aid for CREST, the Japan Science and Technology Agency (JST). The computational resource was provided by Research Institute for Information Technology in the Kyushu University. This work was also conducted as a Joint Usage / Research Center for Interdisciplinary Large-Scale Information Infrastructures in Japan and Advanced Computational Scientific Program 2017, 2018 Research Institute for Information Technology, Kyushu University.
Publisher Copyright:
© 2018 IEEE.
PY - 2019/1/22
Y1 - 2019/1/22
N2 - Presently the power consumption of supercomputer system becomes a critical issue to develop the exascale supercomputer system. On the other hand, the power consumption character of applications is not so considered by the applications developers because their main interest is how fast to run their applications. In this study, we examine and evaluate the power consumption behavior of our Magnetohydrodynamic simulation code which solves the planetary magnetosphere under the constrained power of CPU and DRAM on the x86 computer system. As the results, we found there are some regions in the simulation code which decrease the calculation performance or do not affect the performance under the power capping. This indicates the capability of power optimization without performance degradation using the dynamic power capping in running the application. In addition, we obtained the specific power consumption combinations between CPU and DRAM which greatly affect the calculation performance.
AB - Presently the power consumption of supercomputer system becomes a critical issue to develop the exascale supercomputer system. On the other hand, the power consumption character of applications is not so considered by the applications developers because their main interest is how fast to run their applications. In this study, we examine and evaluate the power consumption behavior of our Magnetohydrodynamic simulation code which solves the planetary magnetosphere under the constrained power of CPU and DRAM on the x86 computer system. As the results, we found there are some regions in the simulation code which decrease the calculation performance or do not affect the performance under the power capping. This indicates the capability of power optimization without performance degradation using the dynamic power capping in running the application. In addition, we obtained the specific power consumption combinations between CPU and DRAM which greatly affect the calculation performance.
UR - http://www.scopus.com/inward/record.url?scp=85062549137&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85062549137&partnerID=8YFLogxK
U2 - 10.1109/HPCC/SmartCity/DSS.2018.00113
DO - 10.1109/HPCC/SmartCity/DSS.2018.00113
M3 - Conference contribution
AN - SCOPUS:85062549137
T3 - Proceedings - 20th International Conference on High Performance Computing and Communications, 16th International Conference on Smart City and 4th International Conference on Data Science and Systems, HPCC/SmartCity/DSS 2018
SP - 626
EP - 631
BT - Proceedings - 20th International Conference on High Performance Computing and Communications, 16th International Conference on Smart City and 4th International Conference on Data Science and Systems, HPCC/SmartCity/DSS 2018
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 20th International Conference on High Performance Computing and Communications, 16th IEEE International Conference on Smart City and 4th IEEE International Conference on Data Science and Systems, HPCC/SmartCity/DSS 2018
Y2 - 28 June 2018 through 30 June 2018
ER -