Hello,
This is a newbie question and might have been covered elsewhere on this
list but I have searched and couldn't find the information I am looking
for. Any help I can get or a pointer to webpage will be helpful. 
I am trying to run Amber8 on EM64T with Intel Compilers. I have never
done this before and needed help to understand how to interpret the
benchmark data. The primary purpose of the benchmarks is to understand
how using different interconnects such as GbE, InfiniBand, 10 GigE may
affect the application performance. 
- What number should I be looking at from the benchmark summary report
(attached).?
- What tests are more relevant to a customer who may be looking to
purchase a cluster to running Amber as a primary application?
I am simply using the benchmarks included with the software for these
tests. 
Regards,
Nikhil 
 
../factor_ix/bench.fix.out
|>>>>>>>>PROFILE of Average TIMES>>>>>>>>> 
|    Read coords time           0.06 ( 0.02% of Total)
|                Build the list            12.87 (91.86% of List )
|                Other                      1.14 ( 8.14% of List )
|             List time                 14.01 ( 9.12% of Nonbo)
|                   Short_ene time            82.50 (92.96% of Direc)
|                   Other                      6.25 ( 7.04% of Direc)
|                Direct Ewald time         88.75 (63.53% of Ewald)
|                Adjust Ewald time          0.70 ( 0.50% of Ewald)
|                   Fill Bspline coeffs        2.87 ( 8.71% of Recip)
|                   Fill charge grid           4.64 (14.10% of Recip)
|                   Scalar sum                 3.21 ( 9.77% of Recip)
|                   Grad sum                   6.84 (20.79% of Recip)
|                      FFT communication ti       5.39 (37.29% of FFT t)
|                      Other                      9.06 (62.71% of FFT t)
|                   FFT time                  14.45 (43.94% of Recip)
|                   Other                      0.88 ( 2.68% of Recip)
|                Recip Ewald time          32.89 (23.55% of Ewald)
|                Force Adjust              15.26 (10.93% of Ewald)
|                Virial junk                1.49 ( 1.07% of Ewald)
|                Start sycnronization       0.58 ( 0.42% of Ewald)
|             Ewald time               139.68 (90.88% of Nonbo)
|          Nonbond force            153.70 (96.66% of Force)
|          Bond/Angle/Dihedral        0.59 ( 0.37% of Force)
|          FRC Collect time           3.90 ( 2.45% of Force)
|          Other                      0.81 ( 0.51% of Force)
|       Force time               159.00 (96.43% of Runmd)
|       Shake time                 1.28 ( 0.78% of Runmd)
|       Verlet update time         1.29 ( 0.78% of Runmd)
|       CRD distribute time        3.31 ( 2.01% of Runmd)
|       Other                      0.01 ( 0.01% of Runmd)
|    Runmd Time               164.89 (73.21% of Total)
|    Other                     60.29 (26.77% of Total)
| Total time               225.24 (100.0% of ALL  )
 
Nikhil Kelshikar
Server Virtualization Business Unit
W: (408) 853-7733
M: (813) 789-8353
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Aug 31 2005 - 18:53:01 PDT