Re: [Ngspice-users] Performance problems with Athlon 64 X2?

SourceForge Headquarters 1320 Columbia Street Suite 310 San Diego, CA 92101 +1 (858) 422-6466

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
  <title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Way back (many, many moons ago!) when I worked at MCNC in the CAzM
simulator group, we had a PD regression suite we tried to establish as
a standard; a benchmarking suite of circuits for comparing other
simulators (spice derivatives) to our own.&nbsp; We called it CircuitSim90.&nbsp;
I think Dr. Franc Brglez at NCSU maintained this for a while, but I'm
not sure where it's at today.&nbsp; I still have a copy squirreled away
somewhere.&nbsp; The circuits are old, based on simple MOS1,2,3 models and
it could definitely use updating, but perhaps it could be a start?<br>
-bill-<br>
<br>
gary wrote:
<blockquote cite="mid...@la..." type="cite">
  <pre wrap="">I'd suggest that someone who knows the workings of nspice come up with 
the test circuit, then we run it on a different machines. I'd assume it 
would make sense to limit this to transient analysis, just to reduce the 
scope.

BTW, I don't recall any flags for multithreading.

John Brunhaver wrote:
  </pre>
  <blockquote type="cite">
    <pre wrap="">If your gcc flags are not set to take advantage of your specific 
architecture, you may be leaving performance on the table. 

Actually, the transition from executing 32-bit to executing 64-bit code 
is known in some memory intensive applications to cause a reduction in 
performance due to the increased size of memory addresses and payloads 
(all else being equal).  At least this was a subject another student at 
my University was pursuing casually and had shared with me a few months 
ago.

However, I have never compiled for or simulated on an AMD machine.  I 
haven't run extensive simulations yet to compare the run times of 1 
64-bit Intel core from a quad-core machine to 1 32-bit P4 thread. Or the 
comparison of 4-Cores to 2-Threads.

I would also appreciate anyone's thoughts on this topic.

On 5/10/07, *Bill Richards* &lt;<a class="moz-txt-link-abbreviated" href="mailto:ric...@tb...">ric...@tb...</a> 
<a class="moz-txt-link-rfc2396E" href="mailto:ric...@tb...">&lt;mailto:ric...@tb...&gt;</a>&gt; wrote:

    Hi-
    I recently upgraded my simulation system from an old XP 1600+ box to an
    Athlon 64 X2 3800+.  I grabbed the x86_64 ngspice package via yum, and
    ran some test cases.  To my surprise, mt test cases ran SLOWER on my new
    Athlon 64 box!!  Being somewhat paranoid of binary RPMs, I downloaded
    the sources and necessary libraries and made it all from scratch.  Ran
    the test cases, and again, the ngspice code on my new box yielded
    exactly the same results as my old box, only over 2X slower!!!  Has
    anyone encountered this kind of behavior with ngspice on an Athlon
    64 box?
    I don't have any other benchmarks to compare performance of this machine
    against my now-defunct old machine.  Any help or suggestions would be
    appreciated!
    Thanks,
    -bill-

    -------------------------------------------------------------------------

    This SF.net email is sponsored by DB2 Express
    Download DB2 Express C - the FREE version of DB2 express and take
    control of your XML. No limits. Just data. Click to get it now.
    <a class="moz-txt-link-freetext" href="http://sourceforge.net/powerbar/db2/">http://sourceforge.net/powerbar/db2/</a>
    _______________________________________________
    Ngspice-users mailing list
    <a class="moz-txt-link-abbreviated" href="mailto:Ngs...@li...">Ngs...@li...</a>
    <a class="moz-txt-link-rfc2396E" href="mailto:Ngs...@li...">&lt;mailto:Ngs...@li...&gt;</a>
    <a class="moz-txt-link-freetext" href="https://lists.sourceforge.net/lists/listinfo/ngspice-users">https://lists.sourceforge.net/lists/listinfo/ngspice-users</a>

------------------------------------------------------------------------

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
<a class="moz-txt-link-freetext" href="http://sourceforge.net/powerbar/db2/">http://sourceforge.net/powerbar/db2/</a>

------------------------------------------------------------------------

_______________________________________________
Ngspice-users mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Ngs...@li...">Ngs...@li...</a>
<a class="moz-txt-link-freetext" href="https://lists.sourceforge.net/lists/listinfo/ngspice-users">https://lists.sourceforge.net/lists/listinfo/ngspice-users</a>
    </pre>
  </blockquote>
  <pre wrap=""><!---->
-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
<a class="moz-txt-link-freetext" href="http://sourceforge.net/powerbar/db2/">http://sourceforge.net/powerbar/db2/</a>
_______________________________________________
Ngspice-users mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Ngs...@li...">Ngs...@li...</a>
<a class="moz-txt-link-freetext" href="https://lists.sourceforge.net/lists/listinfo/ngspice-users">https://lists.sourceforge.net/lists/listinfo/ngspice-users</a>
  </pre>
</blockquote>
<br>
</body>
</html>