Communication latency is central to multiprocessor design. This study presents the design principles of the EM-X distributed-memory multiprocessor towards tolerating communication latency. The EM-X overlaps computation with communication for latency tolerance by multithreading. In particular, we present two types of hardware support for remote memory access: (1) priority-based packet scheduling for thread invocation, and (2) direct remote memory access. The priority-based scheduling policy extends a FIFO ordered thread invocation policy to adopt to different computational needs. The direct remote memory access is designed to overlap remote memory operations with thread execution. The 80-processor prototype of EM-X is developed and is operational since December 1995. We execute several programs on the machine and evaluate how the EM-X effectively overlaps computation with communication toward tolerating communication latency for high performance parallel computing.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Yuetsu KODAMA, Hirohumi SAKANE, Mitsuhisa SATO, Hayato YAMANA, Shuichi SAKAI, Yoshinori YAMAGUCHI, "Message-Based Efficient Remote Memory Access on a Highly Parallel Computer EM-X" in IEICE TRANSACTIONS on Information,
vol. E79-D, no. 8, pp. 1065-1071, August 1996, doi: .
Abstract: Communication latency is central to multiprocessor design. This study presents the design principles of the EM-X distributed-memory multiprocessor towards tolerating communication latency. The EM-X overlaps computation with communication for latency tolerance by multithreading. In particular, we present two types of hardware support for remote memory access: (1) priority-based packet scheduling for thread invocation, and (2) direct remote memory access. The priority-based scheduling policy extends a FIFO ordered thread invocation policy to adopt to different computational needs. The direct remote memory access is designed to overlap remote memory operations with thread execution. The 80-processor prototype of EM-X is developed and is operational since December 1995. We execute several programs on the machine and evaluate how the EM-X effectively overlaps computation with communication toward tolerating communication latency for high performance parallel computing.
URL: https://global.ieice.org/en_transactions/information/10.1587/e79-d_8_1065/_p
Copy
@ARTICLE{e79-d_8_1065,
author={Yuetsu KODAMA, Hirohumi SAKANE, Mitsuhisa SATO, Hayato YAMANA, Shuichi SAKAI, Yoshinori YAMAGUCHI, },
journal={IEICE TRANSACTIONS on Information},
title={Message-Based Efficient Remote Memory Access on a Highly Parallel Computer EM-X},
year={1996},
volume={E79-D},
number={8},
pages={1065-1071},
abstract={Communication latency is central to multiprocessor design. This study presents the design principles of the EM-X distributed-memory multiprocessor towards tolerating communication latency. The EM-X overlaps computation with communication for latency tolerance by multithreading. In particular, we present two types of hardware support for remote memory access: (1) priority-based packet scheduling for thread invocation, and (2) direct remote memory access. The priority-based scheduling policy extends a FIFO ordered thread invocation policy to adopt to different computational needs. The direct remote memory access is designed to overlap remote memory operations with thread execution. The 80-processor prototype of EM-X is developed and is operational since December 1995. We execute several programs on the machine and evaluate how the EM-X effectively overlaps computation with communication toward tolerating communication latency for high performance parallel computing.},
keywords={},
doi={},
ISSN={},
month={August},}
Copy
TY - JOUR
TI - Message-Based Efficient Remote Memory Access on a Highly Parallel Computer EM-X
T2 - IEICE TRANSACTIONS on Information
SP - 1065
EP - 1071
AU - Yuetsu KODAMA
AU - Hirohumi SAKANE
AU - Mitsuhisa SATO
AU - Hayato YAMANA
AU - Shuichi SAKAI
AU - Yoshinori YAMAGUCHI
PY - 1996
DO -
JO - IEICE TRANSACTIONS on Information
SN -
VL - E79-D
IS - 8
JA - IEICE TRANSACTIONS on Information
Y1 - August 1996
AB - Communication latency is central to multiprocessor design. This study presents the design principles of the EM-X distributed-memory multiprocessor towards tolerating communication latency. The EM-X overlaps computation with communication for latency tolerance by multithreading. In particular, we present two types of hardware support for remote memory access: (1) priority-based packet scheduling for thread invocation, and (2) direct remote memory access. The priority-based scheduling policy extends a FIFO ordered thread invocation policy to adopt to different computational needs. The direct remote memory access is designed to overlap remote memory operations with thread execution. The 80-processor prototype of EM-X is developed and is operational since December 1995. We execute several programs on the machine and evaluate how the EM-X effectively overlaps computation with communication toward tolerating communication latency for high performance parallel computing.
ER -