Hello
please find below the respond of ATLAS ADC coordinator.
Elena
Begin forwarded message:
> From: Andrej Filipcic <[log in to unmask]>
> Date: 4 February 2014 12:31:16 GMT
> To: Elena Korolkova <[log in to unmask]>
> Cc: Alexei Klimentov <[log in to unmask]>, Alessandro Di Girolamo <[log in to unmask]>
> Subject: Re: atlas position on multicore resources
>
>
> Hi Elena,
>
> There is no wlcg model
> of implementation yet, and this will be done till October, which is too
> late for ATLAS. In the meantime, our guideline is as follows:
> - implement multicore setup if possible. We do not request right now
> from all the sites to do that, but we will shortly present a list of
> sites we would like to see them multicore enabled, i.e. all T1s and big T2s.
> - we must provide 30-40k cores for MC DC14 by April when we start with
> reconstruction in 8-core mode. For now, the current 20k multicore
> enabled cores are roughly sufficient. The main reason to hold this back
> is still significant mc12 production and the performance problem of mc14
> simulation.
> - analysis is not considered for multicore at this point
>
> So, a rough plan:
> - implement multicore now and try to enable ~30% resources.
> - only production will use multicore
> - try to do it transparently without specific node allocations, i.e
> .dynamically.
>
> In the next few days, we will prepare a list of T2s we wish to setup
> mcore, and we will contact cloud supports to help us.
>
>
we welcome any volunteer site to have multicore queue. We need to gain as much
experience with it as we can to provide a generic solution which will
work for everybody.
>
>
> Cheers,
> Andrej
>
>
>
__________________________________________________
Dr Elena Korolkova
Email: [log in to unmask]
Tel.: +44 (0)114 2223553
Fax: +44 (0)114 2223555
Department of Physics and Astronomy
University of Sheffield
Sheffield, S3 7RH, United Kingdom
|