2015-02-11 17:09 GMT+01:00 Cook, Malcolm <[email protected]>:
>
>
> Fotis, I note your example name, "sandybridge", apparently encoded an
> intel processor microarchitecture, NOT the name of a linux distribution
> (such as c6 or c7 for releases of centOS, as proposed). I'm trying to
> understand the implications of possibly needing to support a heterogeneous
> environment having multiple CentOS versions (6.5 and 7.x) on multiple core
> types (sandybridge) and would appreciate any more clarity here. Are you
> possibly suggesting that buildsets for each combination of microprocessor
> and OS version might be appropriate
> (provoking visions of /opt/apps/{sandybridge,Nehalem}/centOS{6,7}/YYYYMMDD
> )
>
>
>
Hi Malcom,
In case it's helpful, my approach to support different cpu types is to have
a different soft stack by cpu type and then use automounter to access the
right software stack.
As example, I have all my soft stacks in this paths:
/mnt/soft/apps/{sandybridge,nehalem,bulldozer}
Then I use aumounter to "map" /soft/apps to the right folder in
/mnt/soft/apps depending on the cpu type of the machine. In all my compute
nodes the path where the software is accesible is /soft/apps (this is what
my users know about) but automounter is mapping /soft/apps to the right
folder in /mnt/soft/apps/{sandybridge,nehalem,bulldozer}
Of course, for this you need to rebuild all your applications for all the
cpu types you want to support and have a different soft stack for each cpu
but once you have an easyconfig building the application 1 time or 100
times is easy :)
I suppose that adapting that approach to also support multiple
distributions shouldn't be difficult.
regards,
Pablo.
--
Pablo Escobar López
HPC systems engineer
Biozentrum, University of Basel
Swiss Institute of Bioinformatics SIB
Email: [email protected]
Phone: +41 61 267 21 80
http://www.biozentrum.unibas.ch
