[slurm-users] Slurm and available libraries

Patrick Goetz pgoetz at math.utexas.edu
Wed Jan 17 09:50:06 MST 2018

On 01/17/2018 08:12 AM, Ole Holm Nielsen wrote:
> John: I would refrain from installing the old default package 
> "environment-modules" from the Linux distribution, since it doesn't seem 
> to be maintained any more.
> Lmod, on the other hand, is actively maintained and solves some problems 
> with the old "environment-modules" software.

Sorry to thread jack, but I have a related lmod question.  My first 
experience with lmod was setting up an OpenHPC cluster, where the module 
command is defined as a function:

[pgoetz at lakers etc]$ type module
module is a function
module ()
     eval $($LMOD_CMD bash "$@");
     [ $? = 0 ] && eval $(${LMOD_SETTARG_CMD:-:} -s sh)

I thought this was standard, but am now working with an entirely 
different software stack in a multiple GPU compute environment.  I was 
shocked to learn that the definition of module doesn't appear to be 
standardized at all.  Here is how it's defined for the CryoEM compute 
cluster at another university:

[pgoetz at fangtooth ~]$ which module
module: 	 aliased to set _prompt="$prompt";set prompt="";eval 
`/opt/qb3/Modules/$MODULE_VERSION/bin/modulecmd tcsh !*`; set 
_exit=$status; set prompt="$_prompt";unset _prompt;; test 0 = $_exit;

Can anyone shed some light on the situation?  I'm very surprised that a 
module script isn't just an explicit command that comes with the lmod 
package, and am curious as to why this isn't completely standard.

More information about the slurm-users mailing list