[slurm-dev] Re: Query about Compute + GPUs

2017-11-03 Thread Merlin Hartley
! Merlin -- Merlin Hartley Computer Officer MRC Mitochondrial Biology Unit Cambridge, CB2 0XY United Kingdom > On 3 Nov 2017, at 15:43, Ing. Gonzalo E. Arroyo > <garr...@ifimar-conicet.gob.ar <mailto:garr...@ifimar-conicet.gob.ar>> wrote: > > Hi Merlin! Thanks for helping.

[slurm-dev] Re: Query about Compute + GPUs

2017-11-03 Thread Merlin Hartley
Sounds like you would need 2 different NodeName lines - one in each partition. -- Merlin Hartley Computer Officer MRC Mitochondrial Biology Unit Cambridge, CB2 0XY United Kingdom > On 3 Nov 2017, at 15:08, Ing. Gonzalo E. Arroyo > <garr...@ifimar-conicet.gob.ar> wrote: > >

[slurm-dev] Re: Running jobs are stopped and reqeued when adding new nodes

2017-10-23 Thread Merlin Hartley
A workaround is to pre-configure future nodes and mark them as down - then when you add them you can just mark them as up. (see the DownNodes parameter) Hope this helps! Merlin -- Merlin Hartley Computer Officer MRC Mitochondrial Biology Unit Cambridge, CB2 0XY United Kingdom > On 22 Oct 2

[slurm-dev] Re: Finding job command after fails

2017-10-16 Thread Merlin Hartley
You could also use a simple epilog script to save the output of ‘scontrol show job’ to a file/database. M -- Merlin Hartley Computer Officer MRC Mitochondrial Biology Unit Cambridge, CB2 0XY United Kingdom > On 15 Oct 2017, at 20:49, Ryan Richholt <ryanrichh...@gmail.com&

[slurm-dev] GPU accounting (WAS Re: sreport not reporting gpu info...)

2017-10-09 Thread Merlin Hartley
user who exclusively uses GPU machines (4 GPUs and 16 CPUs per machine). Any idea what I’ve missed? Thanks Merlin -- Merlin Hartley Computer Officer MRC Mitochondrial Biology Unit Cambridge, CB2 0XY United Kingdom > On 6 Oct 2017, at 20:30, Tim Carlson <tim.s.carl...@gmail.com> wrote

[slurm-dev] Accounting estimates/calculations

2017-09-14 Thread Merlin Hartley
S/gpu=160.0" Many thanks for your time Merlin -- Merlin Hartley Computer Officer MRC Mitochondrial Biology Unit Cambridge, CB2 0XY United Kingdom

[slurm-dev] Re: Slurm query

2017-07-04 Thread Merlin Hartley
That is the default partition … i.e. if the user doesn’t specify a partition, that is where their jobs will run. Makes more sense when you have multiple partitions. Hope this is useful! M -- Merlin Hartley Computer Officer MRC Mitochondrial Biology Unit Cambridge, CB2 0XY United Kingdom

[slurm-dev] Re: Slurm Light?

2017-04-13 Thread Merlin Hartley
= #AccountingStoragePass= #AccountingStorageUser= #AccountingStorageType=accounting_storage/filetxt #AccountingStorageLoc=/var/spool/slurmd/slurmdb.txt Hope this helps! Merlin -- Merlin Hartley Computer Officer MRC Mitochondrial Biology Unit Cambridge, CB2 0XY United Kingdom

[slurm-dev] Re: Slurm.conf configuration

2017-04-03 Thread Merlin Hartley
Hi Suprita You just need multiple NodeName lines, something like this: NodeName=testmaster CPUs=2 NodeName=testclient CPUs=1 PartitionName=debug Nodes=testmaster,testclient Default=YES MaxTime=INFINITE State=UP Hope this helps! Merlin -- Merlin Hartley Computer Officer MRC Mitochondrial

[slurm-dev] Re: Exclusive socket configuration help

2017-03-22 Thread Merlin Hartley
Hi Cyrus I think you should specify the memory requirements in your sbatch script - the default would be to allocate all the memory for a node - thus ‘filling’ it even with a 1 cpu job. #SBATCH --mem 1G Hope this helps! Merlin -- Merlin Hartley Computer Officer MRC Mitochondrial Biology Unit

[slurm-dev] Re: SLURM reports much higher memory usage than really used

2016-12-15 Thread Merlin Hartley
the memory estimate was anything like correct. In the new year I may try accounting again but with "MemLimitEnforce=no” set as well :) Merlin -- Merlin Hartley IT Systems Engineer MRC Mitochondrial Biology Unit Cambridge, CB2 0XY United Kingdom > On 15 Dec 2016, at 10:32, Uwe Sauter <uw