problem in accelerated MD with NAMD-GPU

From: Thomas Evangelidis (tevang3_at_gmail.com)
Date: Mon Nov 19 2012 - 11:19:56 CST

Dear NAMD developers,

This is essentially a continuation of an older thread:

http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l.2011-2012/3855.html

I have used the precompiled NAMD_CVS-2012-11-19_Linux-x86_64-multicore and
NAMD_CVS-2012-11-19_Linux-x86_64-multicore-CUDA binaries. Below are the
"BOND"
"ANGLE", "DIHED", "ELECT", "VDW", "KINETIC", "TOTAL", and "TEMP" values
I got from the .log files with NAMD plot plugin of VMD. Only "DIHED" ,
"KINETIC" and "TEMP" are comparable. It seems that no boost was applied to
the "TOTAL" energy with NAMD-CUDA.

Step CPU GPU

"BOND"
0 556.3471 556.3471
100 824.4590 913.8393
200 786.0528 891.1186
300 722.3652 916.2721
400 691.9713 940.2994
500 699.5470 867.2131
600 647.4432 909.8798
700 678.7098 935.1221
800 718.0975 889.7386
900 643.5344 891.9886
1000 665.8129 993.3422

"ANGLE"
0 1489.0836 1489.0836
100 2290.2081 2279.6028
200 2235.6279 2215.0142
300 2092.6019 2232.4232
400 2021.8599 2230.5256
500 1989.5349 2292.7212
600 1950.1460 2307.8106
700 1913.1973 2409.2568
800 1808.2876 2348.2903
900 1890.0223 2323.0422
1000 1864.1678 2314.5148

"DIHED"
0 2173.2210 2173.2210
100 2736.2255 2779.3166
200 2744.4755 2750.2174
300 2761.2298 2753.5947
400 2737.6699 2733.4879
500 2741.6743 2744.1523
600 2734.6001 2716.4655
700 2743.2592 2727.6693
800 2725.2575 2749.1877
900 2736.3057 2752.6277
1000 2745.8980 2767.1214

"ELECT"
0 -454148.7061 -454148.7086
100 -437451.7727 -453494.0105
200 -429773.3332 -453184.7815
300 -427142.3312 -452975.2804
400 -427325.0289 -453101.1229
500 -427265.1194 -452819.5838
600 -428016.3683 -452654.7390
700 -428787.4667 -452653.6915
800 -428038.2602 -452044.5900
900 -427338.5386 -453043.6885
1000 -427615.1559 -452948.8649

"VDW"
0 66090.9339 66092.7970
100 65287.5139 65714.6218
200 58474.9551 65623.5025
300 56425.7953 65354.2750
400 56733.8768 65591.1033
500 56724.9622 65473.3231
600 57366.9548 65153.0557
700 58300.8767 64915.5734
800 57625.4728 64473.6502
900 56981.5434 65489.9931
1000 57272.6945 65546.4413

"KINETIC"
0 59555.8770 59558.0374
100 57451.3849 58870.1792
200 58241.6037 59124.3286
300 58758.2684 59386.9614
400 58832.9532 59208.1018
500 58540.8742 59022.4185
600 58904.8136 59072.0660
700 58821.1550 59223.1018
800 59112.7730 58982.6400
900 59350.6355 58996.4601
1000 59524.9722 59278.3877

"TOTAL"
0 -324283.2435 -324279.2225
100 -308861.9813 -322936.4508
200 -307290.6181 -322580.6002
300 -306382.0707 -322331.7540
400 -306306.6979 -322397.6049
500 -306568.5268 -322419.7557
600 -306412.4107 -322495.4614
700 -306330.2687 -322442.9681
800 -306048.3718 -322601.0832
900 -305736.4973 -322589.5769
1000 -305541.6106 -322049.0573

"TEMP"
0 297.1301 297.1409
100 286.6306 293.7091
200 290.5731 294.9771
300 293.1508 296.2874
400 293.5234 295.3950
500 292.0662 294.4686
600 293.8819 294.7163
700 293.4645 295.4699
800 294.9194 294.2702
900 296.1061 294.3391
1000 296.9759 295.7457

Let me know if there is other information you might need.

thanks,
Thomas

-- 
======================================================================
Thomas Evangelidis
PhD student
University of Athens
Faculty of Pharmacy
Department of Pharmaceutical Chemistry
Panepistimioupoli-Zografou
157 71 Athens
GREECE
email: tevang_at_pharm.uoa.gr
          tevang3_at_gmail.com
website: https://sites.google.com/site/thomasevangelidishomepage/

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:22:16 CST