archive: RE: SETI @ Home progress

RE: SETI @ Home progress

Smith, S. Jay ( (no email) )
Wed, 19 May 1999 12:18:56 -0400

So, a 65% 'useful' rate reduces the work units to ~ 134K per day....they
need to ramp up processing about 5-fold to get to this level, plus catch up
data stored from january to may....

this logic leads to about 1 million users if they want to process data in
NRT...or 5-fold greater than the current 200K...
or they could speed up the client software to get there....

Its possible they could get to this level, but it might help to encourage
participants by having a monetary award for the 'winner.'

This experiment makes me wonder if companies might use distributed
processing to 'farm out' work and pay for CPU cycles across the internet.
Thus, they could avoid buying 'big' computers....


Jay Smith

> -----Original Message-----
> From: Dan Kennedy [SMTP:dkennedy@acspdx.com]
> Sent: Wednesday, May 19, 1999 11:25 AM
> To: Seti List (E-mail)
> Subject: RE: SETI @ Home progress
>
> [Jay Smith wrote:] ...number of units generated per day would be =
> 24*60*60
> * 256 / 107 = 206.7 K.
>
>
> Somewhere in the SETI@Home site I read that they expected to collect
> useful
> data 65% of the time. That is because the data is only good for us when
> the
> telescope is still. When the carriage is moving to a new location, or
> tracking the sky, the data is not recorded for SETI@Home. The main reason
> is
> the need to use the 12 seconds 'through the beam' timing for gaussian
> fits.
>
> - Dan
>