QCDOC:A 10 teraflops computer for tightly-coupled calculations

P. A. Boyle; D. Chen; N. H. Christ; M. Clark; S. D. Cohen; C. Cristian; Z. Dong; A. Gara; B. Joo; C. Jung; C. Kim; L. Levkova; X. Liao; G. Liu; R. D. Mawhinney; S. Ohta; K. Petrov; T. Wettig; A. Yamaguchi

QCDOC:A 10 teraflops computer for tightly-coupled calculations

P. A. Boyle, D. Chen, N. H. Christ, M. Clark, S. D. Cohen, C. Cristian, Z. Dong, A. Gara, B. Joo, C. Jung, C. Kim, L. Levkova, X. Liao, G. Liu, R. D. Mawhinney, S. Ohta, K. Petrov, T. Wettig, A. Yamaguchi

Advanced Computing for Nuclear, Particle

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

8 Scopus citations

Abstract

Numerical simulations of the strong nuclear force, known as quantum chromodynamics or QCD, have proven to be a demanding, forefront problem in high-performance computing. In. this report, we describe a new computer, QCDOC (QCD On a Chip), designed for optimal price/performance in the study of QCD. QCDOC uses a six-dimensional, low-latencymesh network to connect processing nodes, each of which includes a single custom ASIC, designed by our collaboration and built by IBM, plus DDR SDRAM. Each node has a peak speed of 1 Gigaflops and two 12,288 node, 10+ Teraflops machines are to be completed in the fall of 2004. Currently, a 512 node machine is running, delivering eciencies as high as 45% of peak on the conjugate gradient solvers that dominate our calculations and a 4096-node machine with a cost of $1.6M is under construction. This should give us a price/performance less than $1 per sustained Megaflops.

Original language	English
Title of host publication	IEEE/ACM SC2004 Conference - Bridging Communities, Proceedings
Pages	449-461
Number of pages	13
State	Published - 2004
Event	IEEE/ACM SC2004 Conference - Bridging Communities - Pittsburgh, PA, United States Duration: Nov 6 2004 → Nov 12 2004

Publication series

Name	IEEE/ACM SC2004 Conference, Proceedings

Conference

Conference	IEEE/ACM SC2004 Conference - Bridging Communities
Country/Territory	United States
City	Pittsburgh, PA
Period	11/6/04 → 11/12/04

Cite this

Boyle, P. A., Chen, D., Christ, N. H., Clark, M., Cohen, S. D., Cristian, C., Dong, Z., Gara, A., Joo, B., Jung, C., Kim, C., Levkova, L., Liao, X., Liu, G., Mawhinney, R. D., Ohta, S., Petrov, K., Wettig, T., & Yamaguchi, A. (2004). QCDOC:A 10 teraflops computer for tightly-coupled calculations. In IEEE/ACM SC2004 Conference - Bridging Communities, Proceedings (pp. 449-461). (IEEE/ACM SC2004 Conference, Proceedings).

@inproceedings{23e5b04d0569450ea72888fc5564ba53,

title = "QCDOC:A 10 teraflops computer for tightly-coupled calculations",

abstract = "Numerical simulations of the strong nuclear force, known as quantum chromodynamics or QCD, have proven to be a demanding, forefront problem in high-performance computing. In. this report, we describe a new computer, QCDOC (QCD On a Chip), designed for optimal price/performance in the study of QCD. QCDOC uses a six-dimensional, low-latencymesh network to connect processing nodes, each of which includes a single custom ASIC, designed by our collaboration and built by IBM, plus DDR SDRAM. Each node has a peak speed of 1 Gigaflops and two 12,288 node, 10+ Teraflops machines are to be completed in the fall of 2004. Currently, a 512 node machine is running, delivering eciencies as high as 45% of peak on the conjugate gradient solvers that dominate our calculations and a 4096-node machine with a cost of $1.6M is under construction. This should give us a price/performance less than $1 per sustained Megaflops.",

author = "Boyle, {P. A.} and D. Chen and Christ, {N. H.} and M. Clark and Cohen, {S. D.} and C. Cristian and Z. Dong and A. Gara and B. Joo and C. Jung and C. Kim and L. Levkova and X. Liao and G. Liu and Mawhinney, {R. D.} and S. Ohta and K. Petrov and T. Wettig and A. Yamaguchi",

year = "2004",

language = "English",

isbn = "0769521533",

series = "IEEE/ACM SC2004 Conference, Proceedings",

pages = "449--461",

booktitle = "IEEE/ACM SC2004 Conference - Bridging Communities, Proceedings",

note = "IEEE/ACM SC2004 Conference - Bridging Communities ; Conference date: 06-11-2004 Through 12-11-2004",

}

Boyle, PA, Chen, D, Christ, NH, Clark, M, Cohen, SD, Cristian, C, Dong, Z, Gara, A, Joo, B, Jung, C, Kim, C, Levkova, L, Liao, X, Liu, G, Mawhinney, RD, Ohta, S, Petrov, K, Wettig, T & Yamaguchi, A 2004, QCDOC:A 10 teraflops computer for tightly-coupled calculations. in IEEE/ACM SC2004 Conference - Bridging Communities, Proceedings. IEEE/ACM SC2004 Conference, Proceedings, pp. 449-461, IEEE/ACM SC2004 Conference - Bridging Communities, Pittsburgh, PA, United States, 11/6/04.

TY - GEN

T1 - QCDOC:A 10 teraflops computer for tightly-coupled calculations

AU - Boyle, P. A.

AU - Chen, D.

AU - Christ, N. H.

AU - Clark, M.

AU - Cohen, S. D.

AU - Cristian, C.

AU - Dong, Z.

AU - Gara, A.

AU - Joo, B.

AU - Jung, C.

AU - Kim, C.

AU - Levkova, L.

AU - Liao, X.

AU - Liu, G.

AU - Mawhinney, R. D.

AU - Ohta, S.

AU - Petrov, K.

AU - Wettig, T.

AU - Yamaguchi, A.

PY - 2004

Y1 - 2004

N2 - Numerical simulations of the strong nuclear force, known as quantum chromodynamics or QCD, have proven to be a demanding, forefront problem in high-performance computing. In. this report, we describe a new computer, QCDOC (QCD On a Chip), designed for optimal price/performance in the study of QCD. QCDOC uses a six-dimensional, low-latencymesh network to connect processing nodes, each of which includes a single custom ASIC, designed by our collaboration and built by IBM, plus DDR SDRAM. Each node has a peak speed of 1 Gigaflops and two 12,288 node, 10+ Teraflops machines are to be completed in the fall of 2004. Currently, a 512 node machine is running, delivering eciencies as high as 45% of peak on the conjugate gradient solvers that dominate our calculations and a 4096-node machine with a cost of $1.6M is under construction. This should give us a price/performance less than $1 per sustained Megaflops.

AB - Numerical simulations of the strong nuclear force, known as quantum chromodynamics or QCD, have proven to be a demanding, forefront problem in high-performance computing. In. this report, we describe a new computer, QCDOC (QCD On a Chip), designed for optimal price/performance in the study of QCD. QCDOC uses a six-dimensional, low-latencymesh network to connect processing nodes, each of which includes a single custom ASIC, designed by our collaboration and built by IBM, plus DDR SDRAM. Each node has a peak speed of 1 Gigaflops and two 12,288 node, 10+ Teraflops machines are to be completed in the fall of 2004. Currently, a 512 node machine is running, delivering eciencies as high as 45% of peak on the conjugate gradient solvers that dominate our calculations and a 4096-node machine with a cost of $1.6M is under construction. This should give us a price/performance less than $1 per sustained Megaflops.

UR - http://www.scopus.com/inward/record.url?scp=23944508737&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:23944508737

SN - 0769521533

T3 - IEEE/ACM SC2004 Conference, Proceedings

SP - 449

EP - 461

BT - IEEE/ACM SC2004 Conference - Bridging Communities, Proceedings

T2 - IEEE/ACM SC2004 Conference - Bridging Communities

Y2 - 6 November 2004 through 12 November 2004

ER -

QCDOC:A 10 teraflops computer for tightly-coupled calculations

Abstract

Publication series

Conference

Other files and links

Fingerprint

Cite this