We are researchers at three separate instititions and would like to
pool our spare CPUs (about 100) toward working on a single problem. Our
potential compute nodes all have outgoing HTTP (at least through username
protected web proxies) and we all have the ability to modify files
stored on a web server at a 4th location. None of us have root/admin
access nor incoming access (e.g. ssh nor ftp) from the outside.
After reading the manual and trying some examples it is not clear how
I could set this up other than having the application specific code do the
scheduling and file transfer - which defeats the purpose. Maybe condor-g
can somehow do this?
Can someone recommend a general condor strategy for this?
Thanks! -- ------------------------------------------------------------------------ Jason
Brazile
jbrazile@xxxxxxxxx
![[Computer Systems Lab]](https://www-auth.cs.wisc.edu/pics/csl_logo.gif){kind=logo}