Stata
NOTE: Please contact [[w|hpc-help]] if you are interested in running Stata on the main-campus HPC clusters.
Our license for is limited to six concurrent users and includes the Basic version, Standard version and MP (parallel) version command-line and X11 GUI version. The maximum number of cores that can be used in a single stata calculation is 16.
Stata is currently installed on all main-campus clusters and can be used on any system, we recommend the bora/him or gust clusters.
Stata , therefore, you should only request one node at most to run stata calculations.
Stata can be run in the command-line or through the graphical interface.
Running Stata interactively through the command-line:
>> salloc -N1 -n4 -t30
This would put you on a compute node with 30 min of run time and 4 compute cores. Once on the compute node you can run stata-mp to launch stata).
If you want to use the graphical interface, first you must ensure that you are logged into the cluster with X11 forwarding (see offices/it/services/researchcomputing/using/connecting/index for more information), and then:
>> salloc -N1 -n4 -t30 --x11
Then you can launch the Stata gui using xstata-mp.
30 min and 4 comute cores is only a suggestion, you can run the interactive job up to the max walltime for the cluster and use up to 16 cores. (see: offices/it/services/researchcomputing/atwm/systemarchitecture/nodes/index for more information about the available clusters).
Keep in mind that adding more cores may not speed up your calculation as much as you expect and depends on the functionality used in your Stata use.
Finally, Stata can also run in batch mode.
An example SLURM job script for use on the bora/hima cluster would be:
| #!/bin/tcsh #SBATCH --job-name=stata #SBATCH -N 1 --ntasks-per-node 8 -C hi #SBATCH -t 12:00:00 module load stata stata-mp ... |
This would give your job 8 cores on a hima node to run stata-mp.