inconsistency between "Benchmark" and "Timing" output

From: Vlad Cojocaru (vlad.cojocaru_at_mpi-muenster.mpg.de)
Date: Mon Nov 25 2013 - 09:08:58 CST

• Next message: k.manzoorolajdad: "can i use swap partition to increase ram in namd?"
• Previous message: Axel Kohlmeyer: "Re: diffusion of protein into water box"
• Next in thread: Norman Geist: "AW: inconsistency between "Benchmark" and "Timing" output"
• Reply: Norman Geist: "AW: inconsistency between "Benchmark" and "Timing" output"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]

Dear all,

I am facing some weird performance issue with NAMD (CVS code from
2013-10-02 compiled with intel compiler 13.0 and intel mpi 4.1.0).
Below are the "Benchmark" and "Timing" outputs. You will notice some
weird jumps in performance during some time chuncks ...

I am running on an intel-based cluster with infiniband.

Does anybody have any idea what can cause this ?

Best wishes
Vlad

Info: Benchmark time: 320 CPUs 0.0065802 s/step 0.0507731 days/ns
657.715 MB memory
Info: Benchmark time: 320 CPUs 0.00668338 s/step 0.0515693 days/ns
657.715 MB memory
Info: Benchmark time: 320 CPUs 0.00624472 s/step 0.0481846 days/ns
657.715 MB memory
Info: Benchmark time: 320 CPUs 0.00594146 s/step 0.0458446 days/ns
657.715 MB memory
Info: Benchmark time: 320 CPUs 0.00611934 s/step 0.0472171 days/ns
657.715 MB memory
Info: Benchmark time: 320 CPUs 0.00609806 s/step 0.0470529 days/ns
657.715 MB memory

TIMING: 83046000 CPU: 194.989, 0.00685107/step Wall: 194.989,
0.00685107/step, 0.14844 hours remaining, 665.828125 MB of memory in use.
TIMING: 83048000 CPU: 207.793, 0.0064024/step Wall: 207.793,
0.0064024/step, 0.135162 hours remaining, 665.828125 MB of memory in use.
TIMING: 83050000 CPU: 220.836, 0.00652158/step Wall: 220.836,
0.00652158/step, 0.134055 hours remaining, 665.828125 MB of memory in use.
TIMING: 83052000 CPU: 282.356, 0.0307597/step Wall: 282.356,
0.0307597/step, 0.615194 hours remaining, 665.828125 MB of memory in use.
TIMING: 83054000 CPU: 295.411, 0.00652754/step Wall: 295.411,
0.00652754/step, 0.126924 hours remaining, 665.828125 MB of memory in use.
TIMING: 83056000 CPU: 317.971, 0.0112801/step Wall: 317.971,
0.0112801/step, 0.213069 hours remaining, 665.828125 MB of memory in use.
TIMING: 83058000 CPU: 330.732, 0.00638029/step Wall: 330.732,
0.00638029/step, 0.116972 hours remaining, 665.828125 MB of memory in use.
TIMING: 83060000 CPU: 384.659, 0.0269637/step Wall: 384.659,
0.0269637/step, 0.479355 hours remaining, 665.828125 MB of memory in use.
TIMING: 83062000 CPU: 397.824, 0.0065825/step Wall: 397.824,
0.0065825/step, 0.113365 hours remaining, 666.015625 MB of memory in use.
TIMING: 83064000 CPU: 410.618, 0.00639666/step Wall: 410.618,
0.00639666/step, 0.106611 hours remaining, 666.015625 MB of memory in use.

-- 
Dr. Vlad Cojocaru
Max Planck Institute for Molecular Biomedicine
Department of Cell and Developmental Biology
Röntgenstrasse 20, 48149 Münster, Germany
Tel: +49-251-70365-324; Fax: +49-251-70365-399
Email: vlad.cojocaru[at]mpi-muenster.mpg.de
http://www.mpi-muenster.mpg.de/research/teams/groups/rgcojocaru

• Next message: k.manzoorolajdad: "can i use swap partition to increase ram in namd?"
• Previous message: Axel Kohlmeyer: "Re: diffusion of protein into water box"
• Next in thread: Norman Geist: "AW: inconsistency between "Benchmark" and "Timing" output"
• Reply: Norman Geist: "AW: inconsistency between "Benchmark" and "Timing" output"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:21:57 CST