In fact, Intel uses SLURM as a resource manager for their 48-core
"cluster on a chip":
http://www.hpcwire.com/hpcwire/2009-12-02/intel_unveils_48-core_research_chip.html
You'll just want to configure a simple set of plugins (e.g. no database).
Quoting Brian Gilmer <[email protected]>:
Yes, it is possible. It is also possible to run both slurmctrl and
slurmd on the same computer. There are enterprise rational for using
SLURM like being able to have common resource allocation for all
assets.
On Wed, Oct 26, 2011 at 9:28 AM, Mike Wykes <[email protected]> wrote:
Hi there,
I am a newcomer to the world of gride engines / queing systems and
would like to ask a simple question if I may:
Can I use slurm to manage job execution on a single multi-core
machine? I.e. Can I use a single machine both as the server (where
jobs are sumbitted and the daemons are running) and the sole compute
node?
I would like to be able to queue up jobs utilising varying numbers of
processors and memory and have slurm manage execution.
Perhaps slurm is overkill for this purpose, but I have not found any
free alternatives. Any other suggestions would be much appreciated!
Many thanks,
Mike
--
Speak when you are angry--and you will make the best speech you'll
ever regret.
- Laurence J. Peter
