This shows you the differences between two versions of the page.
Next revision | Previous revision | ||
parallel_environment [2014/02/16 20:30] rdiazgar created |
parallel_environment [2014/02/25 17:35] (current) rdiazgar |
||
---|---|---|---|
Line 1: | Line 1: | ||
====== Parallel Environment ====== | ====== Parallel Environment ====== | ||
- | The cluster has a module that sets a parallel environment. This module can also be used as a way to 'reserve' slots in a desired queue. | + | ==== Multiple slots in a single node ==== |
+ | |||
+ | If we want to run a job that uses multiple CPU's, we have to reserve as many slots as CPU's in the queue. To do that, we need to use the SMP environment: | ||
+ | |||
+ | <code> | ||
+ | qsub -pe smp N | ||
+ | </code> | ||
+ | |||
+ | Where N is the number of slots/CPU's required. Please note that all slots will be reserved in the same node, and if no node has that amount of available slots, the job will wait until that condition is met. | ||
+ | |||
+ | ==== Multiple slots across nodes ==== | ||
+ | |||
+ | The cluster has a module that sets a parallel environment to run across different nodes. In order to do that, please refer to the OpenMPI documentation to code your program accordingly. | ||
To load the module, simply type: | To load the module, simply type: | ||
- | <code>module load openmpi | + | <code> |
+ | module load openmpi | ||
</code> | </code> | ||