Subject: Re: Attaching Climate Change Experiment Deleted SETI
From: f/fgeorge
Date: 29/05/2006, 14:24
Newsgroups: alt.sci.seti

On Sun, 28 May 2006 15:21:57 -0400, "Bob" <lost@puddy.com> wrote:

"Canopus" wrote:
... Then I noticed it was going to take 2887hrs+ to complete the Climate 
Change experiment and as it had preempted SETI, I suspended it and let 
SETI resume.  All's well now.

Well...

Looks like my 1.4 GHz Celeron is going to take 3300 hours to complete a 
"Sulfur" work unit.  You must have a 1.5 GHz Pentium.

They gave me a year to complete the work unit.  Should take less than half a 
year.  No reason not to complete it.  I think I'll make that my deadline to 
build a 3 GHz dual core, 2 MB cache x 2.  Newegg sells the microprocessor 
for $200, without hyperthreading.  Add hyperthreading, and the uP goes up to 
$1000.  See if I can get the new computer running before the present work 
unit completes.

HT does not speed up the process, it can in fact slow it down. I have
an Intel 840 EE, that is a dual core running at 3.2ghz with a 1 meg L2
cache on each core and HT included. I do 4 units at a time. The old
regular Boinc units were taking about an hour to complete 4 units
using an optimized version of Boinc. The new enhanced units are taking
as little as 3 1/2 hours to a bit over 9 hours, using a non optimized
version.
The problem with HT is that the cache is not designed to use it as
efficently as the regular cores. Therefore the HT suffers. A person
using an Intel 840 3.2ghz machine without the HT can do units a little
bit faster than I can. I prefer to do more at a time, but if I turned
HT off I could do them faster.
Just info to help you make the cpu choice. When I bought mine it was
$1000.00 for just the cpu. My wife paid though, I was happily retired
and she couldn't yet and got jealous and found me a job. I bought a
state of the art pc to compensate.