<br><div class="gmail_quote">On Wed, Dec 22, 2010 at 5:48 PM, Satish Balay <span dir="ltr"><<a href="mailto:balay@mcs.anl.gov" target="_blank">balay@mcs.anl.gov</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<div>On Wed, 22 Dec 2010, Yongjun Chen wrote:<br>
<br>
> Satish,<br>
><br>
> I have reconfigured the PETSC with –download-mpich=1 and<br>
> –with-device=ch3:sock. The results show that the speed up can now remain<br>
> increasing when computing cores increase from 1 to 16. However, the maximum<br>
> speed up is still only around 6.0 with 16 cores. The new log files can be<br>
> found in the attachment.<br>
<br>
</div>Perhaps this was mentioned eariler. Performance doesn't scale with<br>
number of cores. [its depends on both scalable compute unites -aka<br>
cores, as well as scalable memory modules]<br>
<br>
When the hardware is not designed to provide scalable performance -<br>
expecting it is wrong. The goal should be to extract max performance<br>
out of a given piece of hardware - not scalable performance. <br></blockquote><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
Wrt with-device=ch3:sock - it might not be the best performer for<br>
shared memory Try the default 'device=ch3:nemsis'<br>
<font color="#888888"><br>
Satish</font></blockquote></div><br><br>
I am now trying with --with-device=ch3:nemsis. Hope it can have a little better performance<br>
<br><br clear="all"><br>