[gmx-users] (no) superlinear Scaling on Opterons

[Jens Krueger]

As a little update I want to share the following numbers. Maybe they are
useful for somebody, who wants to configure his next cluster. :-)

Davids suggestion with the -shuffle hit exactly the crucial point. Thanks.

cores ns/day (speedup)

      -              -shuffle        -shuffle -sort
01    0.289          0.285           0.286
02    0.487 (1.69)   0.567  (1.99)   0.667  (2.33)
04    0.768 (2.66)   1.122  (3.94)   1.313  (4.59)
08    1.330 (4.60)   2.087  (7.32)   2.504  (8.76)
12    1.776 (6.15)   2.861 (10.04)   3.429 (11.99)
16    2.137 (7.39)   3.600 (12.63)   4.320 (15.10)
20    2.550 (8.82)   3.982 (13.97)   4.881 (17.07)


Gromacs-3.3.1 compiled with gcc
mvapich2-9.8-13
rockscluststers-4.3
dual AMD Opteron 2218 (2.6 GHz) (4 cores per node)
Mellanox IB adapter

-- 
Dr. Jens Krueger

Institute of Biophotonics
School of Biomedical Science and Engineering
National Yang-Ming University
155, Li-Non St., Sec. 2
Taipei, 112, Taiwan (R.O.C.)

mobile: 0921452940

http://www.ym.edu.tw/~wfischer

-------------- next part --------------
A non-text attachment was scrubbed...
Name: jk.vcf
Type: text/x-vcard
Size: 146 bytes
Desc: not available
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20071031/7cc01b67/attachment.vcf>