Abstract
Heterogeneous clusters using accelerators are widely used for high-performance computing system. In such systems, the inter-node communication among accelerators becomes bottleneck due to the data transfer between the accelerator and the host. To eliminate this overhead, we have been developing a novel communication system realizing direct communication among accelerators over computation nodes under the HA-PACS (Highly Accelerated Parallel Advanced system for Computational Sciences) project. Also we are investigating high-level parallel programming language, and several practical application programs on our concept, as well as studying the enhancement of TCA and developing system software stack in the CREST project.
| Original language | English |
|---|---|
| Title of host publication | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
| Publisher | Springer Verlag |
| Pages | 463-474 |
| Number of pages | 12 |
| Volume | 9040 |
| ISBN (Print) | 9783319162133 |
| DOIs | |
| Publication status | Published - 2015 |
| Event | 11th International Symposium on Applied Reconfigurable Computing, ARC 2015 - Bochum, Germany Duration: 2015 Apr 13 → 2015 Apr 17 |
Publication series
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Volume | 9040 |
| ISSN (Print) | 03029743 |
| ISSN (Electronic) | 16113349 |
Other
| Other | 11th International Symposium on Applied Reconfigurable Computing, ARC 2015 |
|---|---|
| Country | Germany |
| City | Bochum |
| Period | 15/4/13 → 15/4/17 |
Fingerprint
Keywords
- Accelerator computing
- GPU direct
- Interconnection network
- PCI express
- Remote DMA
ASJC Scopus subject areas
- Computer Science(all)
- Theoretical Computer Science
Cite this
Towards unification of accelerated computing and interconnection for extreme-scale computing. / Hanawa, Toshihiro; Kodama, Yuetsu; Boku, Taisuke; Amano, Hideharu; Murai, Hitoshi; Umemura, Masayuki; Sato, Mitsuhisa.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9040 Springer Verlag, 2015. p. 463-474 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9040).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Towards unification of accelerated computing and interconnection for extreme-scale computing
AU - Hanawa, Toshihiro
AU - Kodama, Yuetsu
AU - Boku, Taisuke
AU - Amano, Hideharu
AU - Murai, Hitoshi
AU - Umemura, Masayuki
AU - Sato, Mitsuhisa
PY - 2015
Y1 - 2015
N2 - Heterogeneous clusters using accelerators are widely used for high-performance computing system. In such systems, the inter-node communication among accelerators becomes bottleneck due to the data transfer between the accelerator and the host. To eliminate this overhead, we have been developing a novel communication system realizing direct communication among accelerators over computation nodes under the HA-PACS (Highly Accelerated Parallel Advanced system for Computational Sciences) project. Also we are investigating high-level parallel programming language, and several practical application programs on our concept, as well as studying the enhancement of TCA and developing system software stack in the CREST project.
AB - Heterogeneous clusters using accelerators are widely used for high-performance computing system. In such systems, the inter-node communication among accelerators becomes bottleneck due to the data transfer between the accelerator and the host. To eliminate this overhead, we have been developing a novel communication system realizing direct communication among accelerators over computation nodes under the HA-PACS (Highly Accelerated Parallel Advanced system for Computational Sciences) project. Also we are investigating high-level parallel programming language, and several practical application programs on our concept, as well as studying the enhancement of TCA and developing system software stack in the CREST project.
KW - Accelerator computing
KW - GPU direct
KW - Interconnection network
KW - PCI express
KW - Remote DMA
UR - http://www.scopus.com/inward/record.url?scp=84926661910&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84926661910&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-16214-0_43
DO - 10.1007/978-3-319-16214-0_43
M3 - Conference contribution
AN - SCOPUS:84926661910
SN - 9783319162133
VL - 9040
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 463
EP - 474
BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PB - Springer Verlag
ER -