0,000,000 for cluster computing

[John Brookes]

I agree with Alan that claims should be taken with the _appropriate_ pinch
of salt, but reliability is _so_ important that I think that
misinterpretation shouldn't be an issue. Whilst - for farms, at least - it
is sometimes OK to refer to the system as a whole, when it comes to HP(T)C,
the mean time to _node_ failure is more than significant. If you're running
a single job on, say, 32 nodes that'll last 10 days (not entirely typical,
but far from ridiculous) and one of the nodes fails after 9 days.......
Oops! That'll be a fair amount of cash down the drain then.

John Brookes
Quadrics
T: +44 (0)117 9155500
F: +44 (0)117 9075395
E: johnb at quadrics.com
W3: www.quadrics.com


> -----Original Message-----
> From: Alan Scheinine [mailto:scheinin at crs4.it]
> Sent: 12 November 2002 13:16
> To: beowulf at beowulf.org
> Subject: Re: 0,000,000 for cluster computing
> 
> 
> This message uses a character set that is not supported by 
> the Internet Service.  To view the original message content,  
> open the attached message. If the text doesn't display 
> correctly, save the attachment to disk, and then open it 
> using a viewer that can display the original character set. 
>