Article ID: 667 - Last Modified: June 7, 2011
When I tried to divide a Glide job into 10 subjobs and used 10 CPUs to run it, only 7 of the 10 passed, while 3 died. Is there a maximum number that a job can be divided into?
You should be able to divide Glide jobs into hundreds of subjobs without difficulty, as long as the number of ligands per subjob isn't too small (which would make the distribution inefficient). It is also a good idea to create multiple subjobs per CPU, so that the subjobs aren't too large.
The number of CPUs you can use (i.e., the number of simultaneous subjobs) will be limited by the number of Glide licenses you have available. License inavailability will cause subjobs to die. Some queuing systems can support license checking, by which jobs are held in the queue until licenses become available. Please see Section 6.4 of the Installation Guide for more information.
Apart from license inavailability, subjobs can die due to system problems, such as the network or the hardware. You should check in the log file for the job or the failed subjobs to see what kind of failure occurred.
Keywords: multiple CPUs, subjobs, Glide
Type the words or phrases on which you would like to search, or click here to view a list of all
Knowledge Base articles