[Beowulf] SGI to offer Windows on clusters

[John Hearns]

Robert G. Brown wrote:
> 
> http://atlas-proj-computing-tdr.web.cern.ch/atlas-proj-computing-tdr/Html/Computing-TDR.htm 
 >
> its drivers in a modern distro and this "we haven't ported or tested
> past RH 7.3" mentality that may have finally disappeared with e.g.
> Scientific Linux representing a much more up to date and portable and
> rebuildable version of much of e.g. the CERN code base and other tools,
> but I suspect that part of it is still there.
> 
> ATLAS members on the list?  Comments?
> 

As a former high energy physicist, I can say that you have hit so many 
nails on the head there.
The comment on the software stack is spot on. Roll on the day I say when 
  it is common on a grid to package a Xen or VMware image of your 
environment to go with the complicated/hard to compile code. Especially 
on job farms, I do not see how this can impact performance much.

The issue of "I've bought the latest and greatest Intel/Nvidia chipset 
motherboard, but SL v3 doesn't run on it out of the box" is a common 
feature on the SL mailing list. I have real sympathy with the folks at 
Femilab and CERN who try to answer these questions. Especially from the 
ones re "SL doesn't have drivers for the wireless on my laptop" - on a 
distro built and maintained for huge job farms.



And re. the future version of Scientific Linux, there has been debate on 
the list re. co-operating with CENTos and essentially using CENTos as a 
base, and SL being an overlay of specific application and library RPMs.
Pros and cons either way there.


John Hearns