GPUs: An Oasis in the Supercomputing Desert - INSPIRE

DataBETA

GPUs: An Oasis in the Supercomputing Desert

Waseem Kamleh
(
- Adelaide U.
)

Dec, 2012

7 pages

Part of Proceedings, 30th International Symposium on Lattice Field Theory (Lattice 2012) : Cairns, Australia, June 24-29, 2012

Published in:

PoS LATTICE2012 (2012) 186

Contribution to:

- Lattice 2012

e-Print:

1212.4573 [hep-lat]

DOI:

10.22323/1.164.0186

Report number:

ADP-12-52-T819

View in:

ADS Abstract Service

reference search2 citations

Citations per year

Abstract: (arXiv)

A novel metric is introduced to compare the supercomputing resources available to academic researchers on a national basis. Data from the supercomputing Top 500 and the top 500 universities in the Academic Ranking of World Universities (ARWU) are combined to form the proposed '500/500' score for a given country. Australia scores poorly in the 500/500 metric when compared with other countries with a similar ARWU ranking, an indication that HPC-based researchers in Australia are at a relative disadvantage with respect to their overseas competitors. For HPC problems where single precision is sufficient, commodity GPUs provide a cost-effective means of quenching the computational thirst of otherwise parched Lattice practitioners traversing the Australian supercomputing desert. We explore some of the more difficult terrain in single precision territory, finding that BiCGStab is unreliable in single precision at large lattice sizes. We test the CGNE and CGNR forms of the conjugate gradient method on the normal equations. Both CGNE and a modified form of CGNR (with restarts) provide reliable convergence for quark propagator calculations in single precision.

Note:

Proceedings of the 30th International Symposium on Lattice Field Theory, June 24-29, 2012, Cairns, Australia, PoS(Lattice 2012)186

quark: propagator
lattice field theory
quenching
microprocessor: graphics
numerical calculations
lattice
numerical methods

References(6)

Figures(1)

[1]

Building the International Lattice Data Grid

et al.

- Comput.Phys.Commun. 182 (2011) 1208-1214
•
e-Print:
- 0910.1692
•
DOI:
- 10.1016/j.cpc.2011.01.027

[2]

2+1 Flavor Lattice QCD toward the Physical Point

PACS-CS

Collaboration

•

S. Aoki
(
- Tsukuba U., GSPAS and
- RIKEN BNL
)

et al.

- Phys.Rev.D 79 (2009) 034503
•
e-Print:
- 0807.1661
•
DOI:
- 10.1103/PhysRevD.79.034503

[3]

Lattice QCD as a video game

Gyozo I. Egri
(
- Eotvos U.
)
,
Zoltan Fodor
(
- Eotvos U. and
- Wuppertal U. and
- UC, San Diego
)
,
Christian Hoelbling
(
- Wuppertal U.
)
,
Sandor D. Katz
(
- Eotvos U. and
- Wuppertal U.
)
,
Daniel Nogradi
(
- Wuppertal U.
)

et al.

- Comput.Phys.Commun. 177 (2007) 631-639,
- Comput.Phys.Commun. 177 (2007) 631-639
•
e-Print:
- hep-lat/0611022
•
DOI:
- 10.1016/j.cpc.2007.06.005

[4]

Solving Lattice QCD systems of equations using mixed precision solvers on GPUs

QUDA

Collaboration

•

M.A. Clark
(
- Harvard-Smithsonian Ctr. Astrophys. and
- Harvard U.
)

et al.

- Comput.Phys.Commun. 181 (2010) 1517-1528
•
e-Print:
- 0911.3191
•
DOI:
- 10.1016/j.cpc.2010.05.002

[5]

H. A. van der Vorst, Bi-CGSTAB: and Smoothly Converging Variant of Bi-CG for the Solution of Nonsymmetric Linear Systems J. Sci and

A. Fast

- Stat.Comput. 13 (1991) 631-644

[6]

Methods of conjugate gradients for solving linear systems

- J.Res.Natl.Bur.Stand. 49 (1952) 6, 409
•
DOI:
- 10.6028/jres.049.044