{"id":103,"date":"2020-01-22T19:39:04","date_gmt":"2020-01-22T19:39:04","guid":{"rendered":"https:\/\/ni.cmu.edu\/computing\/?post_type=ht_kb&p=103"},"modified":"2020-02-18T19:28:38","modified_gmt":"2020-02-18T19:28:38","slug":"matlab","status":"publish","type":"ht_kb","link":"https:\/\/ni.cmu.edu\/computing\/knowledge-base\/matlab\/","title":{"rendered":"Matlab"},"content":{"rendered":"

Matlab\u00a0<\/a> is a numerical computing and programming environment with a broad range of functionality (matrix manipulation, numerical linear algebra, general-purpose graphics, etc.). Additionally, special application areas are served by a large number of optional toolboxes.<\/p>\n

Running Matlab on the headnode<\/h4>\n

Users need to avoid using the head node for processing using Matlab or other software. However, Matlab Distributed Computing toolbox\/server is installed on the cluster and this requires running Matlab on the head node due to how it integrates with SLURM.<\/p>\n

How do I run Matlab graphically on the cluster?<\/h4>\n

X Window System (X11) server will be needed on your local computer.\u00a0 How you get this will depend on your local computer’s operating system and your preferences.\u00a0 This is covered in more detail in the article Logging onto the Cluster.<\/a><\/p>\n

Users who want to use Matlab graphically and display it on their local computers will need to schedule an interactive session in SLURM using a similar command to this. \u00a0 In is important to include –x11 and adjust the request to the necessary resources (queue, # cpus, memory and time ) you will need for your interactive session.<\/p>\n

srun --x11 -p cpu --cpus-per-task=1 --mem=10GB --time=4:00:00 --pty $SHELL<\/pre>\n
\"\"<\/p>\n
The following is the command set and output from the above screen capture:<\/p>\n
dpane@my-vm:~$ ssh -Y dpane@mind.cs.cmu.edu\r\ndpane@mind.cs.cmu.edu's password: \r\nLast login: Thu Feb 13 07:38:02 2020 from claymore.cnbc.cmu.edu\r\n.\r\n.\r\n[dpane@mind ~]$ \r\n[dpane@mind ~]$ srun --x11 -p cpu --cpus-per-task=1 --mem=10GB --time=4:00:00 --pty $SHELL\r\n[dpane@mind-1-21-1 ~]$ module load matlab-9.5 \r\n[dpane@mind-1-21-1 ~]$ matlab\r\nMATLAB is selecting SOFTWARE OPENGL rendering.\r\nGtk-Message: 13:05:49.875: Failed to load module \"canberra-gtk-module\"\r\nexit\r\n[dpane@mind-1-21-1 ~]$ exit\r\nexit\r\nsrun: error: mind-1-21-1: task 0: Exited with exit code 1\r\n[dpane@mind ~]$ \r\n<\/pre>\n
If you receive “srun: error: No DISPLAY variable set, cannot setup x11 forwarding.<\/strong>” error, try logging onto the cluster using the X11 option<\/p>\n
[dpane@mind ~]$ srun –x11 -p cpu –cpus-per-task=1 –mem=10GB –time=4:00:00 –pty $SHELL
\nsrun: error: No DISPLAY variable set, cannot setup x11 forwarding.<\/p>\n
You should try using on of the following options when you remote login program using ssh.<\/p>\n
-Y<\/strong> ==>\u00a0 Enables trusted X11 forwarding. Trusted X11 forwardings are not subjected to the X11 SECURITY extension controls.<\/p>\n
Non-interactive Matlab Sessions<\/h4>\n
Using Matlab in the interactive mode, as we explain above, can be used to take advantage of the integrated environment. However, it can also be run in a non-interactive way that is suitable for “batch processing” and will allow users to takefull advantage of cluster resources. In most cases, the preferred mode of operation for Matlab on our cluster is non-interactive.<\/p>\n
One way to run Matlab non-interactively is through<\/p>\n