Preface
vii
Germany, and France, had been conducted to address fundamental research on the
various aspects of HPC software during 2016–2018 (phase II). The project “MUST
Correctness Checking for YML and XMP Programs (MYX)” had been selected.
The correctness checking program, MUST is applied to the XcalableMP program.
XcalableMP was taken as a parallel programming language project in FLAGSHIP 2020 project which was to develop the Japanese flagship supercomputer,
Fugaku, for improving the productivity and performance of parallel programming.
XcalableMP is now available on Fugaku and the performance is enhanced by the
Fugaku interconnect, Tofu-D. Chapter “XcalableMP 2.0 and Future Directions”
presents the current status of XcalableMP on Fugaku and a proposal for the next
version, XcalableMP 2.0. I conclude this book with challenges for future PGAS
models and some retrospectives about XcalableMP.
I hope this book will provide useful information to the XcalableMP programmers. Furthermore, for future exascale computing and beyond, an important role
of programming model is to bridge the gap between a new architecture and
programmer’s view as well as providing a comprehensive architectural view of the
system. I believe that the PGAS model and its extensions will be still an important
and suggestive model integrating computation and memory.
Finally, I would like to thank the members of the XcalableMP Specification
Working Group and all the people who contributed to the XcalableMP project.
Kobe, Japan
Mitsuhisa Sato
July 2020
vii
Germany, and France, had been conducted to address fundamental research on the
various aspects of HPC software during 2016–2018 (phase II). The project “MUST
Correctness Checking for YML and XMP Programs (MYX)” had been selected.
The correctness checking program, MUST is applied to the XcalableMP program.
XcalableMP was taken as a parallel programming language project in FLAGSHIP 2020 project which was to develop the Japanese flagship supercomputer,
Fugaku, for improving the productivity and performance of parallel programming.
XcalableMP is now available on Fugaku and the performance is enhanced by the
Fugaku interconnect, Tofu-D. Chapter “XcalableMP 2.0 and Future Directions”
presents the current status of XcalableMP on Fugaku and a proposal for the next
version, XcalableMP 2.0. I conclude this book with challenges for future PGAS
models and some retrospectives about XcalableMP.
I hope this book will provide useful information to the XcalableMP programmers. Furthermore, for future exascale computing and beyond, an important role
of programming model is to bridge the gap between a new architecture and
programmer’s view as well as providing a comprehensive architectural view of the
system. I believe that the PGAS model and its extensions will be still an important
and suggestive model integrating computation and memory.
Finally, I would like to thank the members of the XcalableMP Specification
Working Group and all the people who contributed to the XcalableMP project.
Kobe, Japan
Mitsuhisa Sato
July 2020