array jobs accessing one binary in population genetics software

Question

[prasundutta87]

Hi,

I am trying to run the program SNeP (https://sourceforge.net/projects/snepnetrends/files/binaries/). I have the binary installed in my home directory of the SGE cluster I am using. I wish to run multiple instances of the program for different sets of populations. I have 10 different populations and I have a shell script that runs an array job that would run SNeP on the 10 respective populations.

I want to know if it is all right for my 10 jobs to access the single binary file of SNeP. Will it overload the usage of the local copy of SNeP and slow down the program run?

Regards, Prasun

Answer 1

[ATpoint]

That will work just fine, don't worry. I/O bottlenecks can happen if many jobs start reading or writing to disk even though a good cluster with a decent parallel file system must/should also be able to handle this.

[prasundutta87]

Thanks @ATpoint..that's reassuring..the program is in generally quite slow and also hogs a lot of memory..was trying to see if the binary access is a bottleneck that can be avoided..had also thought of making multiple copies of the binary and then use it for each population..