[Beowulf] value of parallel programming experience (was: Checkpointing using flash)
diep at xs4all.nl
Tue Sep 25 10:04:51 PDT 2012
On Sep 25, 2012, at 5:44 PM, Hearns, John wrote:
> It is not so much about parallel programming experience but about
> scientific software development career path. Quite often parallel
> skills are needed anyway. A former colleague and a good friend of mine
> explains it quite nicely here:
> Igor, that's a very good article.
> I must say though that in particle physics there is a job title of
> Physicist Programmer.
You are correct, one of the government organisations here sometimes
has an empty spot for that.
They have their own clusters and aren't connected to the 'grid' AFAIK
with those, which makes them
move easier around.
job requirement : PHD or at least master degree and very good in
job payment: 1800-2200 euro a month (before tax).
Less than half what the average programmer makes building a website.
I would argue that good programming the new generation manycore
hardware is pretty important.
Right now i'm finishing an EGTB generator (computer chess) that's
going to do at a single box 8 core Xeon
in 6 months the same thing like earlier this year 8000+ cores from
Lomonosov took many months.
Achieving something and achieving the same thing in a more effective
algorithmic manner is 2 different fields of
Over here the scientific institutes do not have a budget seemingly to
pay good programmers, and honestely if they had,
odds are there they'd give it to their biggest friend rather than
effectively use such budget what it's needed for.
What's coming at us is not making things easier. The programming
component gets more important i'd argue than the hardware
component. Provided of course you have *some* budget for hardware.
> "The lack of career security often makes talented software
> developers move to industry"
> $ME looks up over the cubicle divider at a large expanse of
> industrial grey desks...
> (not that I'm talented, I hasten to add)
> The contents of this email are confidential and for the exclusive
> use of the intended recipient. If you receive this email in error
> you should not copy it, retransmit it, use it or disclose its
> contents but should return it to the sender immediately and delete
> your copy.
> Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin
> To change your subscription (digest mode or unsubscribe) visit
More information about the Beowulf