1. Home
  2. Research
  3. Centers & Labs
  4. RIKEN Center for Computational Science
  5. Flagship 2020 Project

RIKEN Center for Computational Science Co-Design Team

Team Leader: Junichiro Makino (Ph.D.)

Research Summary

Junichiro  Makino(Ph.D.)

The mission of the Co-design team is to organize the "co-design" between the hardware and software of the exascale system. It is unpractical to design the many-core complex processor of today without taking into account the requirement of applications. At the same time, it is also unrealistic to develop applications without taking into account the characteristics of the processors on which it will run. The term "Co-design" means we will modify both hardware and software to resolve bottlenecks and achieve best performance.

In the early phase, the team lead the effort to estimate the performance of key applications on the new machine, find out the bottlenecks in both the hardware and software sides, and propose improvements.

In the late phase, the team will study the interaction between the hardware, system software, and applications, and investigate how the performance can be improved.

In addition, the team takes into account the direction of the evolution of processors, study the new algorithms which can better utilized the new hardware, and develop DSL (domain-specific language) and frameworks for application development, which help the implementation of new algorithms.

Main Research Fields

  • Informatics

Related Research Fields

  • Complex Systems
  • Interdisciplinary Science & Engineering

Keywords

  • Large Scale Astrophysical Simulation
  • Computer Architecture
  • High Performance Computing

Selected Publications

  • 1.Muranushi, T., Nitadori, K., and Makino, J.:
    "Nanoblock Unroll: Towards the Automatic Generation of Stencil Codes with the Optimal Performance"
    Proceedings of the Second Workshop on Optimizing Stencil Computations, 49-55 (2014).
  • 2.Yoshifuji, N., Sakamoto, R., Nitadori, K., and Makino, J.:
    "Implementation and evaluation of data-compression algorithms for irregular-grid iterative methods on the PEZY-SC processor"
    2016 6th Workshop on Irregular Applications: Architecture and Algorithms (IA3), 58-61 (2016)
  • 3.Muranushi, T., Hotta, H., Makino, J., Nishizawa, S., Tomita, H., Nitadori, K., Iwasawa, M., Hosono, N., Maruyama, Y., Inoue, H., Yashiro, H. and Nakamura, Y.:
    "Simulations of below-ground dynamics of fungi: 1.184 pflops attained by automated generation and autotuning of temporal blocking codes"
    IEEE Press, P.3 13, (2016)
  • 4.Muranushi, T., Nishizawa, S., Tomita, H., Nitadori, K., Iwasawa, M., Maruyama Y., Yashiro, H., Nakamura, Y., Hotta, H., Makino, J., Hosono, N. and Inoue, H.:
    “Automatic generation of efficient codes from mathematical descriptions of stencil computation”
    ACM, 17-22 22, (2016)
  • 5.Sumi, T., Maruyama, Y. and Mitsutake, A.:
    "A reference-modified density functional theory: An application to solvation free-energy calculations for a Lennard-Jones solution"
    The Journal of Chemical Physics, Volume 144, Issue 22, P 224104 (2016)
  • 6.Tanaka, H., Ishikawa, Y., Sakamoto, R., Nakamura, T., Kimura, Y., Nitadori, K., Tsubouchi, M., Makino, J.:
    "Automatic Generation of High-Order Finite-Difference Code with Temporal Blocking For Extreme-Scale Many-Core Systems"
    Fourth International Workshop on Extreme Scale Programming Models and Middleware (2018).

Related Links

Lab Members

Principal investigator

Junichiro Makino
Team Leader

Core members

Daisuke Namekata
Research Scientist
Miyuki Tsubouchi
Technical Staff I

Contact Information

R401 Research Building,
7-1-26 Minatojima-minamimachi, Chuo-ku
Kobe, Hyogo
650-0047 Japan
Tel: +81-(0)78-940-5632
Fax: +81-(0)78-304-4970

Email: jmakino [at] riken.jp

Top