<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:st1="urn:schemas-microsoft-com:office:smarttags" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<meta http-equiv=Content-Type content="text/html; charset=iso-8859-1">
<meta name=Generator content="Microsoft Word 11 (filtered medium)">
<!--[if !mso]>
<style>
v\:* {behavior:url(#default#VML);}
o\:* {behavior:url(#default#VML);}
w\:* {behavior:url(#default#VML);}
.shape {behavior:url(#default#VML);}
</style>
<![endif]--><o:SmartTagType
namespaceuri="urn:schemas-microsoft-com:office:smarttags" name="PersonName"/>
<!--[if !mso]>
<style>
st1\:*{behavior:url(#default#ieooui) }
</style>
<![endif]-->
<style>
<!--
/* Font Definitions */
@font-face
{font-family:Tahoma;
panose-1:2 11 6 4 3 5 4 4 2 4;}
@font-face
{font-family:CMR10;
panose-1:0 0 0 0 0 0 0 0 0 0;}
@font-face
{font-family:CMTT10;
panose-1:0 0 0 0 0 0 0 0 0 0;}
@font-face
{font-family:CMTI10;
panose-1:0 0 0 0 0 0 0 0 0 0;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
{margin:0in;
margin-bottom:.0001pt;
font-size:12.0pt;
font-family:"Times New Roman";}
a:link, span.MsoHyperlink
{color:blue;
text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
{color:purple;
text-decoration:underline;}
span.EmailStyle17
{mso-style-type:personal-reply;
font-family:Arial;
color:navy;}
@page Section1
{size:8.5in 11.0in;
margin:1.0in 1.25in 1.0in 1.25in;}
div.Section1
{page:Section1;}
-->
</style>
</head>
<body lang=EN-US link=blue vlink=purple>
<div class=Section1>
<p class=MsoNormal><font size=2 color=navy face=Arial><span style='font-size:
10.0pt;font-family:Arial;color:navy'>What are you using the machine file for if
you’re using MPD? Are you booting and closing your MPD ring for each
job?<o:p></o:p></span></font></p>
<p class=MsoNormal><font size=2 color=navy face=Arial><span style='font-size:
10.0pt;font-family:Arial;color:navy'><o:p> </o:p></span></font></p>
<p class=MsoNormal><font size=2 color=navy face=Arial><span style='font-size:
10.0pt;font-family:Arial;color:navy'>Quoted from the user manual section on
MPD:<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>If you are using the </span></font><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10'>mpd </span></font><font
size=2 face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>process
manager, which is the default, then many<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>options are available. If you are
using </span></font><font size=2 face=CMTT10><span style='font-size:11.0pt;
font-family:CMTT10'>mpd</span></font><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>, then before you run </span></font><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10'>mpiexec</span></font><font
size=2 face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>,<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>you will have started, or will have
had started for you, a ring of processes<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>called </span></font><font size=2
face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10'>mpd</span></font><font
size=2 face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>’s
(multi-purpose daemons), each running on its own host. It is<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>likely, but not necessary, that each
</span></font><font size=2 face=CMTT10><span style='font-size:11.0pt;
font-family:CMTT10'>mpd </span></font><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>will be running on a separate host.<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>You can find out what this ring of
hosts consists of by running the program<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMTT10><span
style='font-size:11.0pt;font-family:CMTT10'>mpdtrace</span></font><font size=2
face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>. One of the </span></font><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10'>mpd</span></font><font
size=2 face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>’s
will be running on the “local” machine, the one<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>where you will run </span></font><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10'>mpiexec</span></font><font
size=2 face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>. The
default placement of MPI processes, if one<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>runs<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'><o:p> </o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><i><font size=2 face=CMTT10><span
style='font-size:11.0pt;font-family:CMTT10;font-style:italic'>mpiexec -n 10
a.out<o:p></o:p></span></font></i></p>
<p class=MsoNormal style='text-autospace:none'><i><font size=2 face=CMTT10><span
style='font-size:11.0pt;font-family:CMTT10;font-style:italic'><o:p> </o:p></span></font></i></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>is to start the first MPI process
(rank 0) on the local machine and then to<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>distribute the rest around the </span></font><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10'>mpd </span></font><font
size=2 face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>ring one at
a time. If there are more<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>processes than </span></font><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10'>mpd</span></font><font
size=2 face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>’s,
then wraparound occurs. If there are more </span></font><font size=2
face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10'>mpd</span></font><font
size=2 face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>’s
than<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>MPI processes, then some </span></font><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10'>mpd</span></font><font
size=2 face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>’s
will not run MPI processes. Thus any<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>number of processes can be run on a
ring of any size. While one is doing<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>development, it is handy to run only
one </span></font><font size=2 face=CMTT10><span style='font-size:11.0pt;
font-family:CMTT10'>mpd</span></font><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>, on the local machine. Then<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>all the MPI processes will run
locally as well.<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'><o:p> </o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>The first modification to this
default behavior is the </span></font><font size=2 face=CMTT10><span
style='font-size:11.0pt;font-family:CMTT10'>-1 </span></font><font size=2
face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>option to </span></font><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10'>mpiexec<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>(not a great argument name). If </span></font><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10'>-1 </span></font><font
size=2 face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>is
specified, as in<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'><o:p> </o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><i><font size=2 face=CMTT10><span
style='font-size:11.0pt;font-family:CMTT10;font-style:italic'>mpiexec -1 -n 10
a.out<o:p></o:p></span></font></i></p>
<p class=MsoNormal style='text-autospace:none'><i><font size=2 face=CMTT10><span
style='font-size:11.0pt;font-family:CMTT10;font-style:italic'><o:p> </o:p></span></font></i></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>then the first application process
will be started by the first </span></font><font size=2 face=CMTT10><span
style='font-size:11.0pt;font-family:CMTT10'>mpd </span></font><font size=2
face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>in the ring<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><i><font size=2 face=CMTI10><span
style='font-size:11.0pt;font-family:CMTI10;font-style:italic'>after </span></font></i><font
size=2 face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>the local
host. (If there is only one </span></font><font size=2 face=CMTT10><span
style='font-size:11.0pt;font-family:CMTT10'>mpd </span></font><font size=2
face=CMR10><span style='font-size:11.0pt;font-family:CMR10'>in the ring, then
this will be on<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>the local host.) This option is for
use when a cluster of compute nodes has<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>a “head node” where
commands like </span></font><font size=2 face=CMTT10><span style='font-size:
11.0pt;font-family:CMTT10'>mpiexec </span></font><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>are run but not application<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'>processes.<o:p></o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10'><o:p> </o:p></span></font></p>
<p class=MsoNormal style='text-autospace:none'><b><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10;font-weight:bold'>If an </span></font></b><b><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10;font-weight:
bold'>mpd </span></font></b><b><font size=2 face=CMR10><span style='font-size:
11.0pt;font-family:CMR10;font-weight:bold'>is started with the </span></font></b><b><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10;font-weight:
bold'>--ncpus </span></font></b><b><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10;font-weight:bold'>option, then when
it is its turn to<o:p></o:p></span></font></b></p>
<p class=MsoNormal style='text-autospace:none'><b><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10;font-weight:bold'>start a process, it
will start several application processes rather than just<o:p></o:p></span></font></b></p>
<p class=MsoNormal style='text-autospace:none'><b><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10;font-weight:bold'>one before handing
off the task of starting more processes to the next </span></font></b><b><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10;font-weight:
bold'>mpd<o:p></o:p></span></font></b></p>
<p class=MsoNormal style='text-autospace:none'><b><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10;font-weight:bold'>in the ring.</span></font></b><font
size=2 face=CMR10><span style='font-size:11.0pt;font-family:CMR10'> <b><span
style='font-weight:bold'>For example, if the </span></b></span></font><b><font
size=2 face=CMTT10><span style='font-size:11.0pt;font-family:CMTT10;font-weight:
bold'>mpd </span></font></b><b><font size=2 face=CMR10><span style='font-size:
11.0pt;font-family:CMR10;font-weight:bold'>is started with<o:p></o:p></span></font></b></p>
<p class=MsoNormal style='text-autospace:none'><b><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10;font-weight:bold'><o:p> </o:p></span></font></b></p>
<p class=MsoNormal style='text-autospace:none'><b><i><font size=2 face=CMTT10><span
style='font-size:11.0pt;font-family:CMTT10;font-weight:bold;font-style:italic'>mpd
--ncpus=4<o:p></o:p></span></font></i></b></p>
<p class=MsoNormal style='text-autospace:none'><b><i><font size=2 face=CMTT10><span
style='font-size:11.0pt;font-family:CMTT10;font-weight:bold;font-style:italic'><o:p> </o:p></span></font></i></b></p>
<p class=MsoNormal style='text-autospace:none'><b><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10;font-weight:bold'>then it will start
as many as four application processes, with consecutive<o:p></o:p></span></font></b></p>
<p class=MsoNormal style='text-autospace:none'><b><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10;font-weight:bold'>ranks, when it is
its turn to start processes. This option is for use in clusters<o:p></o:p></span></font></b></p>
<p class=MsoNormal style='text-autospace:none'><b><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10;font-weight:bold'>of SMP’s,
when the user would like consecutive ranks to appear on the same<o:p></o:p></span></font></b></p>
<p class=MsoNormal style='text-autospace:none'><b><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10;font-weight:bold'>machine. (In the
default case, the same number of processes might well run<o:p></o:p></span></font></b></p>
<p class=MsoNormal style='text-autospace:none'><b><font size=2 face=CMR10><span
style='font-size:11.0pt;font-family:CMR10;font-weight:bold'>on the machine, but
their ranks would be different.)</span></font></b><b><font size=2 face=CMR10><span
style='font-size:10.0pt;font-family:CMR10;font-weight:bold'><o:p></o:p></span></font></b></p>
<p class=MsoNormal><font size=2 color=navy face=Arial><span style='font-size:
10.0pt;font-family:Arial;color:navy'><o:p> </o:p></span></font></p>
<p class=MsoNormal><font size=2 color=navy face=Arial><span style='font-size:
10.0pt;font-family:Arial;color:navy'>It seems like you should only see the
behavior you are seeing if you did start your MPDs with the --ncpus option. Otherwise,
it should rotate between machines roundrobin style. But I don’t
understand why you’ve got a machine file if you’re using an MPD
ring.<o:p></o:p></span></font></p>
<p class=MsoNormal><font size=2 color=navy face=Arial><span style='font-size:
10.0pt;font-family:Arial;color:navy'><o:p> </o:p></span></font></p>
<p class=MsoNormal><font size=2 color=navy face=Arial><span style='font-size:
10.0pt;font-family:Arial;color:navy'>-Matt Chambers<o:p></o:p></span></font></p>
<p class=MsoNormal><font size=2 color=navy face=Arial><span style='font-size:
10.0pt;font-family:Arial;color:navy'><o:p> </o:p></span></font></p>
<div style='border:none;border-left:solid blue 1.5pt;padding:0in 0in 0in 4.0pt'>
<div>
<div class=MsoNormal align=center style='text-align:center'><font size=3
face="Times New Roman"><span style='font-size:12.0pt'>
<hr size=2 width="100%" align=center tabindex=-1>
</span></font></div>
<p class=MsoNormal><b><font size=2 face=Tahoma><span style='font-size:10.0pt;
font-family:Tahoma;font-weight:bold'>From:</span></font></b><font size=2
face=Tahoma><span style='font-size:10.0pt;font-family:Tahoma'> <st1:PersonName
w:st="on">owner-mpich-discuss@mcs.anl.gov</st1:PersonName> [mailto:<st1:PersonName
w:st="on">owner-mpich-discuss@mcs.anl.gov</st1:PersonName>] <b><span
style='font-weight:bold'>On Behalf Of </span></b>Christian M. Probst<br>
<b><span style='font-weight:bold'>Sent:</span></b> Saturday, May 12, 2007 11:50
PM<br>
<b><span style='font-weight:bold'>To:</span></b> mpich-discuss@mcs.anl.gov<br>
<b><span style='font-weight:bold'>Subject:</span></b> [MPICH] MPICH2 doesn´t
distribute jobs when running applications</span></font><o:p></o:p></p>
</div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'><o:p> </o:p></span></font></p>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'>Hi, folks.<o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'> <o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'>I am running MPICH2 in two servers with 8 cores each one... I have
configured both servers properly and passed all troubleshooting steps provided
in the installation guide.<o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'> <o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'>But when I try to run my applications of interest (both bioinformatics,
mpiClustal and mpiHMMer), the following scenario appears:<o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'> <o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'>If I run just in one server, the job is distributed sucessfully for the
8 cores...<o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'><br>
If I run using both server in my machine file, but using -np 8, all jobs are
distributed to one server, and I have the same time of the previous running
(Ok, expected)... But no job is distributed for the second machine... <o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'> <o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'>If I run using both server in my machine file, with any -np from 9 to 16,
all processes are distributed to just one server (mpdtrace -l appears with
several 0 in the beginning of the line) and I have no results after hours
waiting... <o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'> <o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'>As I said before, if I run the tests, it distributes properly for both
servers, no matter what was set with -np.<o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'> <o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'>Any clues?<o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'> <o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'>Thanks in advance.<o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'>Christian<o:p></o:p></span></font></p>
</div>
<div>
<p class=MsoNormal><font size=3 face="Times New Roman"><span style='font-size:
12.0pt'> <o:p></o:p></span></font></p>
</div>
</div>
</div>
</body>
</html>