Overview of the QCDSP and QCDOC computers

被引：45

作者：

Boyle, PA

Chen, D

Christ, NH

Clark, MA

Cohen, SD

Cristian, C

Dong, Z

Gara, A

Joó, B

Jung, C

Kim, C

Levkova, LA

Liao, X

Liu, G

Mawhinney, RD

Ohta, S

Petrov, K

Wettig, T

Yamaguchi, A

机构：

[1] Univ Edinburgh, Sch Phys, Edinburgh EH9 3JZ, Midlothian, Scotland

[2] Columbia Univ, Dept Phys, New York, NY 10027 USA

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

[4] Brookhaven Natl Lab, Upton, NY 11973 USA

[5] Indiana Univ, Bloomington, IN 47401 USA

[6] US Equity Derivat, Citigrp Quantitat Equity Trading, New York, NY 10013 USA

[7] Univ Regensburg, Inst Theoret Phys, D-93040 Regensburg, Germany

[8] UBS Investment Bank, Stamford, CT 06902 USA

[9] Univ Glasgow, Dept Phys & Astron, Glasgow G12 8QQ, Lanark, Scotland

[10] Gluon Capital LLC, New York, NY 10013 USA

[11] High Energy Accelerator Res Org, Inst Particle & Nucl Studies, Tsukuba, Ibaraki 3050801, Japan

来源：

IBM JOURNAL OF RESEARCH AND DEVELOPMENT | 2005年 / 49卷 / 2-3期

关键词：

D O I：

10.1147/rd.492.0351

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

The QCDSP and QCDOC computers are two generations of midtithousand-node multidimensional mesh-based computers designed to study quantum chromodynamics (QCD), the theory of the strong nuclear force. QCDSP (QCD on digital signal processors). a four-dimensional mesh machine, was completed in 1998; in that year, it won the Gordon Bell prize in the price/performance categorly. Two large installations - of 8,192 and 12,288 nodes, with a combined peak speed of one teraflops-have been in operation since. QCD-on-a-chip (QCDOC) utilizes a sixdimensional mesh and compute nodes fabricated with IBM system-on-a-chip technology. It offers a tenfold impropement in price/performace . Currently 100-node versions are operating, and there are plans to build three 12,288-node, 10-teraflops machines. In this paper, we describe the architecture of both the QCDSP and QCDOC machines, the operating systems employed, the user software environment, and the performance of our application-lattice QCD.

引用

页码：351 / 365

页数：15

共 11 条

[1] Status of the apeNEXT project [J].

Ammendola, R ;

Bodin, F ;

Boucaud, P ;

Cabibbo, N ;

Di Carlo, F ;

De Pietri, R ;

Di Renzo, F ;

Errico, W ;

Fucci, A ;

Guagnelli, M ;

Kaldass, H ;

Lonardo, A ;

de Luca, S ;

Micheli, J ;

Morenas, V ;

Pene, O ;

Petronzio, R ;

Palombi, F ;

Pleiter, D ;

Paschedag, N ;

Rapuano, F ;

De Riso, P ;

Salamon, A ;

Salina, G ;

Sartori, L ;

Schifano, F ;

Simma, H ;

Tripiccione, R .

NUCLEAR PHYSICS B-PROCEEDINGS SUPPLEMENTS, 2003, 119 :1038-1040

[2] A 0.5 TERAFLOPS MACHINE OPTIMIZED FOR LATTICE QCD [J].

ARSENIN, I ;

CHEN, D ;

CHRIST, N ;

EDWARDS, R ;

GARA, A ;

HANSEN, S ;

KENNEDY, A ;

MAWHINNEY, R ;

PARSONS, J ;

SEXTON, J .

NUCLEAR PHYSICS B, 1994, :820-822

[3]

BOYLE PA, 2003, P C COMP HIGH EN PHY, P1

[4]

BOYLE PA, 2003, P 21 INT S LATT FIEL, P838

[5] QCDOC: A 10-teraflops scale computer for lattice QCD [J].

Chen, D ;

Christ, NH ;

Cristian, C ;

Dong, Z ;

Gara, A ;

Garg, K ;

Joo, B ;

Kim, C ;

Levkova, L ;

Liao, X ;

Mawhinney, RD ;

Ohta, S ;

Wettig, T .

NUCLEAR PHYSICS B-PROCEEDINGS SUPPLEMENTS, 2001, 94 :825-832

[6]

Christ N. H., 1989, Nuclear Physics B, Proceedings Supplements, V9, P549, DOI 10.1016/0920-5632(89)90160-6

[7] Recent development and perspectives of machines for lattice QCD [J].

Lippert, T .

NUCLEAR PHYSICS B-PROCEEDINGS SUPPLEMENTS, 2004, 129 :88-101

[8] The 1 Teraflops QCDSP computer [J].

Mawhinney, RD .

PARALLEL COMPUTING, 1999, 25 (10-11) :1281-1296

[9]

PERRINE KA, 2002, P SPIE C VIS DAT AN, P196

[10]

Tripiccione R., 1990, Nuclear Physics B, Proceedings Supplements, V17, P137, DOI 10.1016/0920-5632(90)90227-L

← 1 2 →