Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

How to use mpirun command with infiniband?

Hi there,I just installed CESM on our cluster, and trying to learn how to use it.There is a problem: our cluster support both ethernet (if i was not wrong) and inifiniband two methods, so how should I make changes to my own script or mpirun command to use inifiniband which could complete my job more fast? I google but found only one method, that to set mpirun command in $CASE.run as " mpirun --mca btl self,openib", but it comes along with "Bus error (signal 7)".I'm not familiar with this, so does any one know how to do this? Thank you very much !
 

jedwards

CSEG and Liaisons
Staff member
This is very much a system dependent issue and has nothing to do with CESM.  You need to discuss with your system administrators.qre
 
Hi jedwards, thanks for your reply.  I have discussed with our system administrators. Out cluster do support infiniband, but I need to add some parameter in my script so CESM can work with it. My problem is I'm not sure how to let CESM use this infiniband feature. If I use the original script, i don't know which one (ethernet or infiniband) CESM is using?  Thanks again! 
 
And when I change "#PBS -l" to  "#PBS -l nodes=4:ppn=16:infiniband", the error shows "job exceeds queue resource limits MSG=cannot locate feasible nodes"...
 
Top