@@ -10,7 +10,7 @@ This is the CNMS's newest computing resource.
We *own* ~**2400** cores. Many users must share these, think before you submit.
* Run test jobs to establish timing and parallelization parameters.
* Try to make your walltimes tight. (not always 48:00:00).
* Try to make your walltimes tight. (**not always 48:00:00**).
* Do not flood the queue with jobs, if you have many small jobs batch them. Ask how.
* We will actively thwart gaming the scheduling policies.
...
...
@@ -63,11 +63,15 @@ Most plentiful node is Haswell 2x16, i.e. hw32
** MOAB Torque Cluster **
## Job Submission
There is only one non-experimental queue: batch
### queues
There are now two **queues** for cnms jobs.
* batch
* high_mem
You want to use the qos: **std**
Unless you have a short development job i.e. a test or debug run, then qos:**devel**
### Quality of service (QOS)
* std - generally this is what you want
* devel - short debug, build and experimental runs
* burst - premptable jobs that run on unused CONDO resources (you must request access from @epd)
If you need to run wide relatively short jobs, are experiencing long waits for std and can deal with them being occassionally prempted (i.e. killed) you can request access to qos: **burst** via [XCAMS](https://xcams.ornl.gov/xcams/groups/cades-cnms-burst)
...
...
@@ -76,10 +80,8 @@ This is the obligatory PBS header for a job.
``` shell
#!/bin/bash
#PBS -S /bin/bash
#PBS -m be
#PBS -M youremail@ornl.gov
#PBS -N job_name
#PBS -q batch
#PBS -N <YOUR_JOB_NAME>
#PBS -q <QUEUE_NAME>
#PBS -l nodes=2:ppn=32:hw32
#PBS -l walltime=00:30:00
#PBS -l naccesspolicy=singlejob
...
...
@@ -127,15 +129,9 @@ Please read these, you can waste a great deal of resources if you do not underst
## Advanced
### Quality of Service (QOS)
### Burst QOS
In theory there are two QOS levels useable on Cades:
***std** -- allows use of up to our purchased amount of nodes.
***burst** -- allows use of more nodes but the job can be preempted by std QOS jobs.
