Dear All,
Does anybody know the details of scratch disk management by RELION.
Does RELION clear the scratch disk after a run is finished? Or it just
relies on overwriting the data by a new job?
What happens if the job crashes?
Does the data need to be cleared always manually?
We are sharing GPU nods among more users and scratch disk space
management becomes an issue.
Extending the topic; is there any option one can write the data on scratch just once a then read
it from multiple jobs? For instance when one continues a 3D refinement.
In our case (slow primary data storage/traffic) this option would safe a
lot of computation time. We are experiencing that copying particles to the scratch sometime
takes much more time then the computation itself.
Even worse, while RELION is copying, the allocated computation power is
idling.
Any advice is welcome
Best wishes
Tomas
--
Dr Tomas Kouba
EMBL Grenoble
71 avenue des Martyrs
CS 90181
38042 Grenoble Cedex 9
France
Tel: +33 476 209 478
e-mail: [log in to unmask]
########################################################################
To unsubscribe from the CCPEM list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCPEM&A=1
|