Subject: | |
From: | |
Reply To: | |
Date: | Thu, 7 Sep 2006 09:36:33 -0400 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
Amit;
I found a few more things about MatlabMPI. I am ccing the rescomp
group to see if anyone has any ideas.
1 - the warning about the "type ahead" is being produced by the pause
command. When matlab is launched in graphical display mode I have
no problem with this command, but if I request -display null (or
-nodisplay), then I get zillions of warnings. I solved this in two ways.
1.1 get rid of all pause command in your ".m" files. I do not know
how risky this is
1.2 run matlab in batch mode using : matlab -nojvm -nosplash <
input > output (if desired)
i did not see any problems with 1.2 although it may take a little bit
longer to run (it will be insignificant for long runs).
2 - I have written a few scripts to run matlabmpi which simplifies
things for users. Amit, you can copy these files
from ~combarje/Springs. newmatmpi.scr is the batch script and
my-input.m is the original input file that is to be used to produce RUN.m
3 - I am running your job now and it has produced a few springX.mat files
At 08:19 AM 9/7/2006, you wrote:
>Jaime
>Thank you.
>
>amit
>
>jaime combariza wrote:
>
>>This is what I found.
>>
>>There is an output in the springs.m file that will create large
>>files. i have not been able to find which line(s) is doing it.
>>Then in run_spring.m after you call springs the second time the
>>second line (after)
>>
>> data(i).A=A;
>>
>>gives an error because A is not defined.
>>
>>
>>I ran a test at OSC and I am getting the type-ahead warning as
>>well. It is annoying because it will create big files. I will
>>keep investigating to see if there is a way to get rid of it.
>>
>>
>>
>>At 02:43 PM 9/5/2006, you wrote:
>>
>>>My code in $home/springs/
>>>
>>>gives a warning:
>>>Warning: Type-ahead buffer overflow.
>>>Warning: Type-ahead buffer overflow.
>>>Warning: Type-ahead buffer overflow.
>>>Warning: Type-ahead buffer overflow.
>>>Warning: Type-ahead buffer overflow.
>>>
>>>Then does not create the suitable script to run the files using Matlabmpi.
>>>
>>>Please help. A similar code was working this summer.
>>>Amit
>>>
>>>--
>>>Amit Shukla
>>>Assistant Professor
>>>Department of Manufacturing and Mechanical Engineering
>>>Miami University, Oxford, Ohio 45056
>>>513 529 6043
>>>513 529 1454 (Fax)
>>
>>
>>Jaime E. Combariza, Ph.D.
>>Assistant Director Research Computing
>>http://www.muohio.edu/researchcomputing
>>Miami University
>>(513) 529-5080
>
>--
>Amit Shukla
>Assistant Professor
>Department of Manufacturing and Mechanical Engineering
>Miami University, Oxford, Ohio 45056
>513 529 6043
>513 529 1454 (Fax)
Jaime E. Combariza, Ph.D.
Assistant Director Research Computing
http://www.muohio.edu/researchcomputing
Miami University
(513) 529-5080
|
|
|