|
||||
|
Preparations
Page last edited by Nikolaj Rørbæk Knøsgaard (nirkn) 04/02-2020
For windows usersFollow the instruction on the link below until (but not including) the section "Getting access to the software". https://wiki.fysik.dtu.dk/gpaw/summerschools/summerschool18/setupwin.html MacOS usersFollow the instruction on the link below until (but not including) the section "Log into the databar". https://wiki.fysik.dtu.dk/gpaw/summerschools/summerschool18/setuplinmac.html How to log in to the databarYou use the ssh (Secure SHell) command to create a secure (encrypted) connection to the databar computers. In the terminal, write: How to edit files in the terminalThere are many text editors built in to your terminal and some of them are quite advanced. We recommend that you use the "gedit" text editor for editing files and encourage you to make the following tweak to the gedit editor: - open gedit by entering "gedit" on the command line. This should open a new window with gedit. - Go to menu -> preferences -> editor. Check "insert spaces instead of tabs" and set tab width to 4. These are the standard settings in the Python community. Getting the software(You only have to do this once) We have preinstalled all software for you (ASE and GPAW). To get the sofware open the special initialization file ".bashrc" with gedit on the command line: gedit ~/.bashrc and append the line source ~jjmo/es2020.sh to the end of the file. Now log out and log in again. Submitting jobs to the databar computer clusterMany of the calculations you are going to run will be able to run directly in the terminal. However, you can run in to a situation where the resources required for the calculation are too great and you would then have to submit the job to the databar computer cluster. We have developed a tool that makes this jobs fairly easy (after some initial setup). The tool is called "MyQueue" (link to docs) and can be called using "mq" on the command line. First the setup: Execute the following commands to set up myqueue: mq -h cp ~jjmo/.myqueue/config.py ~/.myqueue/ Now myqueue should be working. Here is a small crash course in using MyQueue (see the documentation for more usage examples): Assume you have a file name script.py. This can then be submitted using MyQueue on 8 cores for 2 hours using the following command: mq submit script.py -R 8:2h To view the status of the submitted jobs execute mq ls |
|||