<div dir="ltr"><div><p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;direction:ltr"><span style="font-size:12pt;font-family:'Times New Roman',serif">Dear Gus and Dear Huiwei </span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;direction:ltr"><span class=""><span style="font-size:12pt;font-family:'Times New Roman',serif"> </span></span><span style="font-size:12pt;font-family:'Times New Roman',serif">I would like to offer my heartfelt appreciation
for your kind consideration.</span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;direction:ltr"><span style="font-size:12pt;font-family:'Times New Roman',serif">After </span><span style="font-size:12pt;font-family:'Times New Roman',serif">Huiwei </span><font face="Times New Roman, serif" size="3">'s last post I am examining
his suggestions. I have installed current version of mpich2 but it could not
solve problem. utilizing same of various values for Ntilei and Ntilej can not solve problem too. Then I check swan model again. Now I am sure it is the reason of errors.
I am trying to sort out what cause those
errors.</font><span lang="FA" dir="RTL" style="font-family:'Times New Roman',serif;font-size:12pt"></span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;direction:ltr"><span style="font-size:12pt;font-family:'Times New Roman',serif"> I find this model mailing<span class=""> </span>list.</span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;direction:ltr"><span style="font-size:12pt;font-family:'Times New Roman',serif">Thanks again</span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;direction:ltr"><span style="font-size:12pt;font-family:'Times New Roman',serif">Cheers</span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt;text-align:justify;direction:ltr"><span style="font-size:12pt;font-family:'Times New Roman',serif">fereshte</span></p><div>
<table cellpadding="0" style="font-family:arial,sans-serif;font-size:13px"><tbody><tr><td style="width:578px"><table cellpadding="0" style="width:578px"><tbody><tr><td><div><br>
</div></td></tr></tbody></table></td></tr></tbody></table></div></div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Mon, Oct 28, 2013 at 8:03 PM, Gus Correa <span dir="ltr"><<a href="mailto:gus@ldeo.columbia.edu" target="_blank">gus@ldeo.columbia.edu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi Fereshteh<br>
<br>
1) I would look for log or error messages in the model<br>
output *before* the final one ("MPI_Abort ...").<br>
<br>
Although it aborted with 2 processors,<br>
that was a graceful termination by the program.<br>
So what is before the last error message may shed some light<br>
on why the model aborted,<br>
and what it requires to run correctly.<br>
<br>
**<br>
<br>
2) MCT is also used for coupling climate models that we use<br>
here. Some of these models it run in SPMD mode (i.e. a single<br>
executable is launched by mpiexec).<br>
Others run in MPMD mode (i.e. several executables<br>
are launched by mpiexec, which requires a mpiexec command<br>
line with more parameters).<br>
<br>
Your mpiexec suggests that the model relies on a<br>
single executable (oceanG).<br>
However, this may/may not be the case, as you mentioned<br>
three separate items also: ROMS, SWAN, and the coupler.<br>
<br>
Is it a single executable or more than one that your model<br>
uses?<br>
[If you describe how you compile the model it may help.]<br>
<br>
**<br>
<br>
3) Are you trying to run it in hybrid mode, i.e.<br>
using both MPI and OpenMP?<br>
<br>
I am not familiar to this model,<br>
so I am just guessing.<br>
<br>
The conventional wisdom is to use parameters names like<br>
"Nthreads" for OpenMP threads, although parameter names are elusive.<br>
OpenMP would add another twist to your processor configuration,<br>
as you would need to provision additional processors<br>
for OpenMP, besides those for MPI.<br>
<br>
**<br>
<br>
4) Does the domain decomposition (Ntilei and Ntilej) have to<br>
be the same for both models?<br>
<br>
<br>
Have you tried something like:<br>
Ntilei=2<br>
Ntilej=2<br>
i.e. 2*2=4 processors for the ocean<br>
plus one processor for SWANN, a total of 5 processors,<br>
then:<br>
<br>
mpiexec -np 5 ...<br>
<br>
In some models we run here (e.g. MITgcm) you need to match precisely<br>
the number of processors to the tiles used to decompose your domain<br>
(and possibly also the OpenMP threads).<br>
<br>
When you tried mpiexec -np 2, did you choose Ntilei=Ntilej=Nthreads=1 for both models?<br>
<br>
Do you have perhaps to add one processor (or more)<br>
specifically for the coupler?<br>
<br>
**<br>
<br>
5) Do your input files (describing the ocean initial state,<br>
and perhaps the wave state), have to be organized based on the<br>
"tile" configuration, or can they span the whole domain?<br>
This could be another source of error.<br>
<br>
6) Does ROMS and SWANN have mailing lists that can perhaps help<br>
you more than the generic MPICH list?<br>
<br>
<br>
I hope this helps,<br>
Gus Correa<div class="im"><br>
<br>
<br>
On 10/27/2013 06:33 AM, fereshteh komijani wrote:<br>
</div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">
Sure.<br>
<br>
ROMS and SWAN models are models for ocean and wave models,<br>
respectively. Coupling of them require 3 input files: <a href="http://coupling_test.in" target="_blank">coupling_test.in</a><br></div>
<<a href="http://coupling_test.in" target="_blank">http://coupling_test.in</a>>, <a href="http://swan_test.in" target="_blank">swan_test.in</a> <<a href="http://swan_test.in" target="_blank">http://swan_test.in</a>> and<br>
<a href="http://roms_test.in" target="_blank">roms_test.in</a> <<a href="http://roms_test.in" target="_blank">http://roms_test.in</a>> which first one is coupled input file<div class="im"><br>
and also one build.bash file which in it user called requested<br>
libraries, cpp optins, header file, compilers (for me gcc and gfortran).<br>
For example in build.bash file I set<br>
<br>
USE_MPI=on = on<br>
USE_MPIF90 =on<br>
<br>
WHICH_MPI=mpich2<br>
<br>
FORT = gfortran<br>
<br>
also with regard that for coupling model, MCT (model coupling toolkit)<br>
is necessary I have installed it and set its include and lib<br>
directories in build.bash file.<br>
<br></div>
In <a href="http://coupling_test.in" target="_blank">coupling_test.in</a> <<a href="http://coupling_test.in" target="_blank">http://coupling_test.in</a>> user call <a href="http://swan_test.in" target="_blank">swan_test.in</a><br>
<<a href="http://swan_test.in" target="_blank">http://swan_test.in</a>>and <a href="http://roms_test.in" target="_blank">roms_test.in</a> <<a href="http://roms_test.in" target="_blank">http://roms_test.in</a>> as wave and<br>
ocean model's inputs.<br>
<br>
In <a href="http://roms_test.in" target="_blank">roms_test.in</a> <<a href="http://roms_test.in" target="_blank">http://roms_test.in</a>> some coefficients, solving<div class="im"><br>
technique and some input forcing file determined.<br>
<br>
for choosing number of nodes for each model there are NtileI and Ntilej<br></div>
in <a href="http://roms_test.in" target="_blank">roms_test.in</a> <<a href="http://roms_test.in" target="_blank">http://roms_test.in</a>> and Nthreads (ocean) and Nthreads<br>
(wave) in <a href="http://coupling_test.in" target="_blank">coupling_test.in</a> <<a href="http://coupling_test.in" target="_blank">http://coupling_test.in</a>> file ( Nthreads<div class="im"><br>
(ocean)=NtileI *Ntilej ) and total nodes are equal to Nthreads<br>
(ocean)+Nthreads (wave).<br>
<br>
whenever i set one node for wave model (SWAN) (Nthreads (wave)=1) after<br>
running by<br>
<br>
mpirun -np 2 ./oceanG <a href="http://coupling_inlet-test.in" target="_blank">coupling_inlet-test.in</a><br></div>
<<a href="http://coupling_inlet-test.in/" target="_blank">http://coupling_inlet-test.<u></u>in/</a>>>mpi.log<div class="im"><br>
<br>
it replies:<br>
<br>
application called MPI_Abort(comm=0x84000002, 4) - process 0<br>
<br></div>
But when setNthreads (wave)>1 (and everything for Nthreads (ocean)) for<div class="im"><br>
example<br>
<br>
mpirun -np 8 ./oceanG <a href="http://coupling_inlet-test.in" target="_blank">coupling_inlet-test.in</a><br></div>
<<a href="http://coupling_inlet-test.in/" target="_blank">http://coupling_inlet-test.<u></u>in/</a>>>mpi.log<div class="im"><br>
<br>
mpi.log file (attaché file) shows that roms model does not have problems<br>
and its nodes are active but nothing happen for swan model after 2 weeks<br>
than running.<br>
<br>
I hope that this information be sufficient.<br>
<br>
All the best<br>
<br>
fereshte<br>
<br>
<br>
<br></div><div class="im">
______________________________<u></u>_________________<br>
discuss mailing list <a href="mailto:discuss@mpich.org" target="_blank">discuss@mpich.org</a><br>
To manage subscription options or unsubscribe:<br>
<a href="https://lists.mpich.org/mailman/listinfo/discuss" target="_blank">https://lists.mpich.org/<u></u>mailman/listinfo/discuss</a><br>
</div></blockquote><div class="HOEnZb"><div class="h5">
<br>
______________________________<u></u>_________________<br>
discuss mailing list <a href="mailto:discuss@mpich.org" target="_blank">discuss@mpich.org</a><br>
To manage subscription options or unsubscribe:<br>
<a href="https://lists.mpich.org/mailman/listinfo/discuss" target="_blank">https://lists.mpich.org/<u></u>mailman/listinfo/discuss</a><br>
</div></div></blockquote></div><br><br clear="all"><div><br></div>-- <br><div style="text-align:center"><font style="background-color:#cc33cc" color="#ffffff" size="4" face="comic sans ms,sans-serif"><strong>**Angel**</strong></font></div>
<div style="text-align:center"> </div>
<div style="text-align:center"><span><strong><u><font style="background-color:#3333ff" color="#ffff33" size="4" face="Arial Narrow"></font></u></strong></span> </div>
</div>