Re: [HTCondor-users] serial jobs in parallel universe

[Alessandra <afortiorama@xxxxxxxxx>]

Hi Greg,

scheduled (by default) in a pure FIFO way.

with by default does it mean it is somehow possible to implement a fair share on the parallel universe?

parallel universe also needs dedicated nodes. Can vanilla universe jobs run there can the parallel universe jobs run spill out?

Incidentally isn't this also a dask setup? I've been wondering for months if dask needs a parallel universe queue or if it can run on a vanilla one (not the case here, but it might be useful to understand this case). Of course if everything mimicks the standard wlcg 8 multicore jobs there is no question but I thought dask could grow to multiple machines.

On Fri, 15 Sept 2023 at 19:59, Greg Thain via HTCondor-users <htcondor-users@xxxxxxxxxxx> wrote:

On 9/14/23 07:10, Alessandra wrote:
> Hi,
>
> I was wondering if anyone here has setup only a parallel universe also
> to run serial jobs. My assumption is that if machine_count=1 and
> cpu_count=1 (or N for multicore jobs) it should work like the vanilla
> universe but without the headache to have dedicated machines or
> clashing shares between different universes when true parallel jobs
> are submitted.
>
> Is this correct? Or are these naive assumptions?


Hi Alessandra:


This will work, but you there is no free lunch. You will lose
fair-share scheduling, as all parallel universe jobs are scheduled (by
default) in a pure FIFO way.

-greg



_______________________________________________
HTCondor-users mailing list
To unsubscribe, send a message to htcondor-users-request@xxxxxxxxxxx with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/htcondor-users/

--

Well you'll still need a tray. (Eddie Izzard)