How to run Jaguar jobs

Don't forget the General Guidelines for running jobs.

Do NOT run jobs on kuaserver. You can either run them on kuanode# or kuaws#.
We currently have a total of 26 Jaguar licenses. This means that there cannot be more than 26 separate Jaguar jobs running in total across our cluster.

Once you have created your Jaguar input file, submit the Jaguar job using the command:

        $SCHRODINGER/jaguar run myfile.in

where myfile.in is the name of your input file; don't forget the $ sign in front.

Jaguar will then generate two files, myfile.log and myfile.out. When the job finishes, it will generate a "restart" file called myfile.01.in.

It's a good idea to take a look at the log file about 2-3 minutes after submitting the job to make sure that it didn't crash because you messed something up in the input file.

Go here to find out more about the log and out files.

When Jaguar runs a job on a client machine, it creates a subdirectory in /usr/tmp/username/ of the client, matching the prefix of the input file you just submitted. (When you submit a job on a node, you can look for it.) Periodically, Jaguar will update the log and out files in your home directory, which is where you will usually view them. If the node stops while your job is running and it is unable to tell the server and update your log and out files appropriately, you can still recover the restart file (called restart.in) from the node and restart your job from where it last left off. Hopefully you won't have to do this often because it usually means a node went down and had to be restarted, or the network went down. (Both are bad!)

To run multiple jobs consecutively, you can use a batch file.