<div>Dear All;</div>
<div> </div>
<div>I am running GATE on a linux cluster. Following the Gate User's manual, I can use the jobsplitter to submit jobs. For instance, after using job splitter such as "gjs -numberofsplits 10 -clusterplatform openmosix
macro.mac", I have 10 mac files under the directory .Gate/macro/. After changing macro.submit to an executable, I can run ./macro.submit under the current directory. </div>
<div> </div>
<div>Normally, it is good to run simulations in the nodes except the login node. This could be done by using "qsub". Could anyone please give me some suggestions about how to run GATE on a cluster efficiently?
</div>
<div> </div>
<div>Thank Jan De Beenhouwer <em>et al. </em>for their paper "Cluster computing software for GATE simulation", Med. Phys. 34 (6), June, 2007.</div>
<div> </div>
<div>Best regards,</div>
<div>Jonathan</div>
<div> </div>