A THROUGHPUT-OPTIMIZED OPTICAL NETWORK FOR DATA-INTENSIVE COMPUTING

被引：61

作者：

Schares, Laurent ^{[1
]}

Lee, Benjamin G. ^{[1
]}

Checconi, Fabio ^{[1
]}

Budd, Russell ^{[1
]}

Rylyakov, Alexander ^{[1
]}

Dupuis, Nicolas ^{[1
]}

Petrini, Fabrizio ^{[1
]}

Schow, Clint L. ^{[1
]}

Fuentes, Pablo ^{[2
]}

Mattes, Oliver ^{[2
]}

Minkenberg, Cyriel ^{[2
]}

机构：

[1] IBM TJ Watson Res Ctr, Yorktown Hts, NY 10598 USA

[2] IBM Res Zurich, Zurich, Switzerland

来源：

IEEE MICRO | 2014年 / 34卷 / 05期

关键词：

Bandwidth; Computer architecture; Data centers; Distributed computing; Graph exploration algorithm; High performance computing; High-speed optical techniques; Networking; Optical fiber networks; Optical network; Optical switches; Optical switching;

D O I：

10.1109/MM.2014.77

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

A NEW CIRCUIT-SWITCHED NETWORK ARCHITECTURE WITH OPTICAL-SWITCH AND BURST-MODE TRANSCEIVER TECHNOLOGY IS PROPOSED TO SUPPORT DEMANDING GRAPH ALGORITHMS IN A DISTRIBUTED-MEMORY SYSTEM. NETWORK SIMULATIONS PREDICT THAT THE SYSTEM COULD ACHIEVE GRAPH PERFORMANCE ON PAR WITH TODAY'S LEADING SUPERCOMPUTERS, AND ITS LIMITED POWER CONSUMPTION WOULD RESULT IN SEVERAL ORDERS OF MAGNITUDE OF EFFICIENCY IMPROVEMENTS THAT COULD ALLOW THE SYSTEM TO FIT WITHIN A FEW RACKS.

引用

页码：52 / 63

页数：12

共 15 条

[1]

Arimilli Baba, 2010, Proceedings of the 18th IEEE Symposium on High Performance Interconnects (HOTI 2010), P75, DOI 10.1109/HOTI.2010.16

[2]

Barker KevinJ., 2005, P ACMIEEE SUPERCOMPU, P16

[3] BUILDING MANY-CORE PROCESSOR-TO-DRAM NETWORKS WITH MONOLITHIC CMOS SILICON PHOTONICS [J].

Batten, Christopher ;

Joshi, Ajay ;

Orcutt, Jason ;

Khilo, Anatol ;

Moss, Benjamin ;

Holzwarth, Charles W. ;

Popovic, Milos A. ;

Li, Hanqing ;

Smith, Henry I. ;

Hoyt, Judy L. ;

Kaertner, Franz X. ;

Ram, Rajeev J. ;

Stojanovic, Vladimir ;

Asanovic, Krste .

IEEE MICRO, 2009, 29 (04) :8-21

[4] Photonic Architectures for High-Performance Data Centers [J].

Beausoleil, Raymond G. ;

McLaren, Moray ;

Jouppi, Norman P. .

IEEE JOURNAL OF SELECTED TOPICS IN QUANTUM ELECTRONICS, 2013, 19 (02)

[5] Traversing Trillions of Edges in Real-time: Graph Exploration on Large-scale Parallel Machines [J].

Checconi, Fabio ;

Petrini, Fabrizio .

2014 IEEE 28TH INTERNATIONAL PARALLEL AND DISTRIBUTED PROCESSING SYMPOSIUM, 2014,

[6] THE IBM BLUE GENE/Q INTERCONNECTION FABRIC [J].

Chen, Dong ;

Eisley, Noel A. ;

Heidelberger, Philip ;

Senger, Robert M. ;

Sugawara, Yutaka ;

Kumar, Sameer ;

Salapura, Valentina ;

Satterfield, David L. ;

Steinmacher-Burow, Burkhard ;

Parker, Jeffrey J. .

IEEE MICRO, 2012, 32 (01) :32-43

[7] Electronic and Photonic Integrated Circuits for Fast Data Center Optical Circuit Switches [J].

Chen, Luis ;

Sohdi, Avantika ;

Bowers, John E. ;

Theogarajan, Luke ;

Roth, Jon ;

Fish, Greg .

IEEE COMMUNICATIONS MAGAZINE, 2013, 51 (09) :53-59

[8]

Cheng Q., 2014, P OPT FIB COMM C

[9] Technology-driven, highly-scalable dragonfly topology [J].

Kim, John ;

Dally, William J. ;

Scott, Steve ;

Abts, Dennis .

ISCA 2008 PROCEEDINGS: 35TH INTERNATIONAL SYMPOSIUM ON COMPUTER ARCHITECTURE, 2008, :77-+

[10]

Koch B. R., 2013, OPT FIB COMM C

← 1 2 →