On 20/09/16 00:14, John DeSantis wrote:
> Nothing like the DRY principle in a config file!
Grin, I guess it is only a single parameter, just that it can have
different values depending on the context. :-)
> All of the times that I've read over the available parameters for
> slurm.conf, this
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256
Chris,
> You can also set that globally.
>
> # Global default for jobs - request 2GB per core wanted.
> DefMemPerCPU=2048
Nothing like the DRY principle in a config file!
All of the times that I've read over the available parameters for
On 19/09/16 20:06, Diego Zuccato wrote:
> I've set the default to 200, so that users are *strongly* encouraged to
> specify the real amount of RAM they need. Or ask for exclusive access to
> a node.
>
> Just a possible hint.
For our systems 2 GB/core is 1/2 of their actual RAM/core ratio on
On 18/09/16 03:45, John DeSantis wrote:
> Try adding a "DefMemPerCPU" statement in your partition definitions, e.g
You can also set that globally.
# Global default for jobs - request 2GB per core wanted.
DefMemPerCPU=2048
All the best,
Chris
--
Christopher SamuelSenior Systems
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256
Balaji,
Try adding a "DefMemPerCPU" statement in your partition definitions, e.g
.:
PartitionName=PY34 Nodes=okdev1368 DefMemPerCPU=512 MaxTime=INFINITE
State=UP shared=force:4
HTH,
John DeSantis
On 09/16/2016 04:44 PM, Balaji Deivam wrote:
>
Thanks Ed, for the quick reply.
I just tried out your settings but it didn't work.
This is our cpu information but it allows only one job to run at a time.
Second job is waiting for resources.
CPU(s):4
On-line CPU(s) list: 0-3
Thread(s) per core:1
Core(s) per socket:1
Hi Balaji –
We had the same experience with a nearly identical version upgrade. We had to
change the SelectTypeParamters value. We also had CR_Core_Memory prior to the
upgrade. We found that for jobs that explicitly specified a memory value, SLURM
would schedule multiple jobs per node. This