Parallel Shell Scripts to utilize Pelican Hosts????

[Vall]

Guys i want to know whether PPSS( http://code.google.com/p/ppss/ ) can be used in PelicanHPC to share the job within the CPU's...I have a simple requirement.. I have a master and a client node...I need to grep through around 10000 files for a particular pattern...My idea is to use clusters to reduce the search time..Meanwhile 'lamexec n0 <program>' didnt work on PelicanHPC 1.7.2...So is there any ideas?? I dont want to go through ther traditional MPI C style Programming ...Plus a simple script should do the trick. ..But i cant get it done...I need ur suggestions Guys!!!

[Vall]

Guys i gone through the script and found that PPSH basically has a weak way of deteting cores..
It uses "cat /proc/cpuinfo | grep "cpu cores" | cut -d ":" -f 2 | uniq | sed -e s/\ //g"
So my question is how can we modify this so as to detect 2 cpu's while using pelican HPC...Any help appreciated

[Vall]

Finally managed to run the Script...it should be 'lamexec n0 sh myscript'...Thanks Mich!!

[kvrrao]

Sir
I want to run octave examples on a few nodes they are running in Octave environment but not with lamexec
as well the pea_example is giving different outputs at different times even though the same two nodes are lamboot'ed.
Any one can helpme in this regard.

KVR Rao

[Michael Creel]

To run pea_example, you first need to run pelican_setup. Then start octave in a terminal on the frontend node, and type "pea_example". You can edit that file to determine how many MPI ranks to use. The output is different every time, because it does not iterate to convergence - a fixed number of iterations is used, and the random numbers are different every time. If you edit the file to iterate to convergence, you should get results that are essentially the same every time, up to the specified tolerance.