Hi,<br><br>Thanks. I have test different cpus. Our institute has two clusters: one is each node has 4 cpu (A), one is one node has only 1 cpu (B). I made different tests on the two clusters and my local computer using the same system. See the following result:
<br><br> A (For 1 hour)<br> # of cpus ; MD steps <br> 4 finished (200000steps for 26:21)<br> 8 finished (200000steps for 40:57)<br> 12 87950<br> 20 42749
<br> 44 5962 !!!!! <br> B (For 1 hour)<br> # of cpu ; MD steps<br> 1 156991 for 56:12<br> 2 179820 <br> 3 200,000 for 54:20
<br> 4 200,000 for 51:12<br> c. Local(single cpu), 200000 steps For 1h52:38<br><br>One can see that <br>1. On cluster A, one nodes(4 cpu) is just as 4 times fast as my local computer.<br>2. More than one nodes will decrease the performancs the gromacs,
<br>3. On cluster B, the more cpu used, the faster gromacs runs. But the difference of speed is not apparent.<br>4. Cluster B with 4 cpus is slow as half as that Cluster A with 1 node (4 cpus)<br><br>I wonder if anyone can tell the bottlenack: the hardware on the cluster or gromacs?
<br><br><br><br><div><span class="gmail_quote">2006/9/14, Mark Abraham <<a href="mailto:mark.abraham@anu.edu.au">mark.abraham@anu.edu.au</a>>:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
> You have to<br>> find your optimum making some tests with your settings. To do that you can<br>> start your simulation and interrupt after a while to have some data logged<br>> in the log file. Then, from the information in that log file you can
<br>> estimate the time that the whole task will take and compare using more or<br>> less number of processors until you find your optimum value.<br><br>Of course, that "while" should be at least of the order of several
<br>minutes. There is a set-up cost borne once at the start of the calculation<br>which is not proportional to the length of the calculation, so you need to<br>run long enough to get out of the time period during which it dominates
<br>the linear component.<br><br>Mark<br><br>_______________________________________________<br>gmx-users mailing list <a href="mailto:gmx-users@gromacs.org">gmx-users@gromacs.org</a><br><a href="http://www.gromacs.org/mailman/listinfo/gmx-users">
http://www.gromacs.org/mailman/listinfo/gmx-users</a><br>Please don't post (un)subscribe requests to the list. Use the<br>www interface or send it to <a href="mailto:gmx-users-request@gromacs.org">gmx-users-request@gromacs.org
</a>.<br>Can't post? Read <a href="http://www.gromacs.org/mailing_lists/users.php">http://www.gromacs.org/mailing_lists/users.php</a><br></blockquote></div><br><br clear="all"><br>-- <br>Sincerely yours,<br>**********************************************
<br>Baofu Qiao, PhD<br>Frankfurt Institute for Advanced Studies<br>**********************************************