<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
Matt you were right<br>
<br>
I was stupidly passing the wrong array (column major order) to
Metis. My code is in Fortran. I am surprised Metis didnt crash
before.<br>
<br>
Thanks anyway.<br>
<br>
Tabrez<br>
<br>
<br>
On 08/25/2011 11:49 PM, Matthew Knepley wrote:
<blockquote
cite="mid:CAMYG4GkLjh9z-TJPqGnXYOFUs-PjTV27OskYBzH2kADzBgSNpw@mail.gmail.com"
type="cite">On Thu, Aug 25, 2011 at 9:15 PM, Tabrez Ali <span
dir="ltr"><<a moz-do-not-send="true"
href="mailto:stali@geology.wisc.edu">stali@geology.wisc.edu</a>></span>
wrote:<br>
<div class="gmail_quote">
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex;">
Hello<br>
<br>
I have an unstructured FE mesh which I am partitioning using
Metis.<br>
<br>
In the first case I only use the element partitioning info and
discard the nodal partitioning info i.e., the original
ordering is same as petsc's global ordering. In the second
case I do use the nodal partitioning info and nodes are
distributed accordingly.<br>
<br>
I would expect that in the 2nd scenario the total number of
MPI messages (at the end of the solve) would be lower than the
1st. However I see that opposite is true. See the plot at <a
moz-do-not-send="true"
href="http://stali.freeshell.org/mpi.png" target="_blank">http://stali.freeshell.org/mpi.png</a><br>
<br>
The number on the y axis is the last column of the "MPI
messages:" field from the -log_summary output.<br>
<br>
Any ideas as to why this is happening. Does relying on total
number of MPI messages as a performance measure even make
sense. Please excuse my ignorance on the subject.<br>
<br>
Alternatively what is a good way to measure how good the Metis
partitioning is?<br>
</blockquote>
<div><br>
</div>
<div>The thing to do here is take a case like 2 proc that can be
completely understood, and get down to the details. I</div>
<div>think there is a probably just a simple misunderstanding
here.</div>
<div><br>
</div>
<div>The first thing to check is that you are partitioning what
you think. By default, Metis partitions the vertices of a
graph,</div>
<div>not elements, Thus you usually have to give Metis the
"dual" of your finite element mesh. I would take a small
(maybe</div>
<div>8 or 10 elements) mesh and look at the original and Metis
partitions. If Metis does not look better, something is wrong.</div>
<div><br>
</div>
<div> Matt</div>
<div> </div>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex;">
Thanks in advance<br>
<font color="#888888">
<br>
Tabrez<br>
<br>
</font></blockquote>
</div>
<br>
<br clear="all">
<div><br>
</div>
-- <br>
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to
which their experiments lead.<br>
-- Norbert Wiener<br>
</blockquote>
<br>
</body>
</html>