An Evaluation of Network Architectures for Next Generation Supercomputers

被引：0

作者：

Chen, Dong ^{[1
]}

Heidelberger, Philip ^{[1
]}

Stunkel, Craig ^{[1
]}

Sugawara, Yutaka ^{[1
]}

Minkenberg, Cyriel ^{[2
]}

Prisacari, Bogdan ^{[3
]}

Rodriguez, German ^{[2
]}

机构：

[1] IBM Corp, Thomas J Watson Res Ctr, Yorktown Hts, NY 10598 USA

[2] Rockley Photon, Zurich, Switzerland

[3] Swiss Fed Inst Technol, Zurich, Switzerland

来源：

PROCEEDINGS OF PMBS 2016: 7TH INTERNATIONAL WORKSHOP ON PERFORMANCE MODELING, BENCHMARKING AND SIMULATION OF HIGH PERFORMANCE COMPUTING SYSTEMS | 2016年

关键词：

Network Architecture; Network Performance; Fat Tree; Dragonfly; Stacked All-to-All; Stacked 2D HyperX;

D O I：

10.1109/PMBS.2016.7

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

We survey network topologies, in particular networks with full all-to-all bandwidth scaling. For more detailed study, we select several recently introduced, promising networks that are cheaper than a 3-level Fat-tree. Through a combination of analysis and simulation on selected supercomputer workloads, we compare these networks according to desirable network properties such as robust performance, low cost, and partitionability. We conclude with observations for future systems.

引用

页码：11 / 21

页数：11

共 50 条

[1] Static Evaluation of a Midimew Connected Torus Network for Next Generation Supercomputers
Al-Shammari, Maryam
Haque, Asrar
Rahman, M. M. Hafizur
[J]. SUSTAINABILITY, 2023, 15 (08)
[2] Research the architectures of the next generation network
Zhang, HJ
Xu, QF
Zhong, YX
[J]. Proceedings of the World Engineers' Convention 2004, Vol A, Network Engineering and Information Society, 2004, : 394 - 397
[3] Backend systems Architectures in the age of the next generation network
Takuji, Tomiyama
[J]. NEC TECHNICAL JOURNAL, 2006, 1 (02): : 51 - 55
[4] RECURSIVE DUAL-NET: A NEW VERSATILE NETWORK FOR SUPERCOMPUTERS OF THE NEXT GENERATION
Li, Yamin
Peng, Shietung
Chu, Wanming
[J]. JOURNAL OF THE CHINESE INSTITUTE OF ENGINEERS, 2009, 32 (07) : 931 - 938
[5] Recursive Dual-Net: A New Universal Network for Supercomputers of the Next Generation
Li, Yamin
Peng, Shietung
Chu, Wanming
[J]. ALGORITHMS AND ARCHITECTURES FOR PARALLEL PROCESSING, PROCEEDINGS, 2009, 5574 : 809 - +
[6] HPCC RandomAccess Benchmark for Next Generation Supercomputers
Aggarwal, Vikas
Sabharwal, Yogish
Garg, Rahul
Heidelberger, Philip
[J]. 2009 IEEE INTERNATIONAL SYMPOSIUM ON PARALLEL & DISTRIBUTED PROCESSING, VOLS 1-5, 2009, : 407 - +
[7] Optical Interconnection Technology for the Next Generation Supercomputers
Sakai, J.
Noda, A.
Yamagishi, M.
Ohtsuka, T.
Sunaga, K.
Sugita, H.
Ono, H.
Yashiki, K.
Kouta, H.
[J]. HIGH PERFORMANCE COMPUTING ON VECTOR SYSTEMS 2009, 2010, : 53 - +
[8] Next Generation Optical-Wireless Converged Network Architectures
Ranaweera, Chathurika
Wong, Elaine
Lim, Christina
Nirmalathas, Ampalavanapillai
[J]. IEEE NETWORK, 2012, 26 (02): : 22 - 27
[9] Next Generation Intra-Vehicle Backbone Network Architectures
Alparslan, Onur
Arakawa, Shin'ichi
Murata, Masayuki
[J]. 2021 IEEE 22ND INTERNATIONAL CONFERENCE ON HIGH PERFORMANCE SWITCHING AND ROUTING (IEEE HPSR), 2021,
[10] Network architectures: Conversion strategies and evolution towards the next-generation network
Schneider, B
[J]. JOURNAL OF THE COMMUNICATIONS NETWORK, 2002, 1 : 47 - 51

← 1 2 3 4 5 →