Re: ::scr Seti@Home & Global Warming

[Ash Argent-Katwala <ash@xxxxxxxxxxx>]

On Fri, 17 May 2002, Dave Turner wrote:
>> My guess is that the power consumption difference isn't going to be that
>> significant
>Even compared to machines that sleep after a period of idle time and then
>dump the memory state to disk and turn off?

Quite hard to spot the machines that that holds for though. It presumes
there aren't any other concurrent processes on those machines. What about
running SETI as a low-priority process on a machine that isn't fully
utilised, but generally fairly busy? If a fair chunk of the overhead is on
running the rest of the system, rather than just the processor, what does
that do for the numbers on multi-processor machines which are being used for
whatever purpose, and the extra processing power is a tiny cost?

As for the grid stuff mentioned later in the thread, I'd be shocked if it
wasn't a case paying more for better service, in which case soaking up the
'spare' cycles with lower priority cheap/free jobs is fair enough. Whether
that bottom rung will really be cheap or free is another matter though.

-- 
ash
a-k
... It's not the heat, it's the humanity