XcalableMP 2.0 and Future Directions
Mitsuhisa Sato, Hitoshi Murai, Masahiro Nakao, Keisuke Tsugane,
Tesuya Odajima, and Jinpil Lee
Abstract This chapter presents the XcalableMP on the Fugaku supercomputer, the
Japanese flagship supercomputer developed by FLAGSHIP2020 project in RIKEN
R-CCS. The porting and the performance evaluation were done as a part of this
project, and the XcalableMP is available for the Fugaku users for improving the
productivity and performance of parallel programing. The performance of XcalableMP on the Fugaku is enhanced by the manycore processor and a new Tofu-D
interconnect. We are now working on the next version, XcalableMP 2.0, for cuttingedge high-performance systems with manycore processors by multithreading and
multi-tasking with integrations of PGAS model and synchronization models. We
conclude this book with retrospectives and challenges for future PGAS models.
1 Introduction
We have been developing a production-level XcalableMP compiler, and make it
available for the K computer’s users as well as the users of conventional clusters.
RIKEN R-CCS has been carrying out the FLAGSHIP 2020 Project [1] to develop
the Japanese flagship supercomputer system following the K computer, the PostK, formally named as “Fugaku” later, since 2014. In the project, XcalableMP was
taken as a parallel programming language project for improving the productivity and
performance of parallel programing. XcalableMP is now available on Fugaku and
the performance is enhanced by the Fugaku interconnect, Tofu-D. The next section
describes the XcalableMP on Fugaku.
M. Sato () · H. Murai · M. Nakao · T. Odajima · J. Lee
RIKEN Center for Computational Science, Kobe, Japan
e-mail: msato@riken.jp; h-murai@riken.jp; masahiro.nakao@riken.jp; tetsuya.odajima@riken.jp;
jinpil.lee@riken.jp
K. Tsugane
Fujitsu Laboratories Ltd., Kawasaki, Kanagawa, Japan
e-mail: tsugane.keisuke@fujitsu.com
© The Author(s) 2021
M. Sato (ed.), XcalableMP PGAS Programming Language,
https://doi.org/10.1007/978-981-15-7683-6_10
245
Mitsuhisa Sato, Hitoshi Murai, Masahiro Nakao, Keisuke Tsugane,
Tesuya Odajima, and Jinpil Lee
Abstract This chapter presents the XcalableMP on the Fugaku supercomputer, the
Japanese flagship supercomputer developed by FLAGSHIP2020 project in RIKEN
R-CCS. The porting and the performance evaluation were done as a part of this
project, and the XcalableMP is available for the Fugaku users for improving the
productivity and performance of parallel programing. The performance of XcalableMP on the Fugaku is enhanced by the manycore processor and a new Tofu-D
interconnect. We are now working on the next version, XcalableMP 2.0, for cuttingedge high-performance systems with manycore processors by multithreading and
multi-tasking with integrations of PGAS model and synchronization models. We
conclude this book with retrospectives and challenges for future PGAS models.
1 Introduction
We have been developing a production-level XcalableMP compiler, and make it
available for the K computer’s users as well as the users of conventional clusters.
RIKEN R-CCS has been carrying out the FLAGSHIP 2020 Project [1] to develop
the Japanese flagship supercomputer system following the K computer, the PostK, formally named as “Fugaku” later, since 2014. In the project, XcalableMP was
taken as a parallel programming language project for improving the productivity and
performance of parallel programing. XcalableMP is now available on Fugaku and
the performance is enhanced by the Fugaku interconnect, Tofu-D. The next section
describes the XcalableMP on Fugaku.
M. Sato () · H. Murai · M. Nakao · T. Odajima · J. Lee
RIKEN Center for Computational Science, Kobe, Japan
e-mail: msato@riken.jp; h-murai@riken.jp; masahiro.nakao@riken.jp; tetsuya.odajima@riken.jp;
jinpil.lee@riken.jp
K. Tsugane
Fujitsu Laboratories Ltd., Kawasaki, Kanagawa, Japan
e-mail: tsugane.keisuke@fujitsu.com
© The Author(s) 2021
M. Sato (ed.), XcalableMP PGAS Programming Language,
https://doi.org/10.1007/978-981-15-7683-6_10
245