Page still under construction - Please forgive the mess...

I have been participating in the Setiathome project since the end of July,1999. I was ranked as high as #20 in late October,1999 before most of my high-powered computer systems went into production. (doing real work!)

I still grab CPU cycles on all of my home machines. I'm not interested in having my name on a list. I do SETI because I believe in the cause. I believe that we are not alone in the Universe. I hope I am the one to discover ET!

Setiahome has achieved truly amazing results. Together, we make up the largest "supercomputer" in the world. No single project has ever used more computing resources.

If you are interested in finding out about the number-crunchers that I used to knock out most of my 40,000+ data sets, read the attached e-mail below. It contains tips for you intel users that might help reduce your dataset times. (nothing you can't find from reading the SETI newsgroups)

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Hi,

Here is a response that I have been sending to people who ask about my
seti results.  I hope this helps answer your questions.  

Let me know if it doesn't.

Thanks,

-tjh

=-=-=-=

There are 2 things to consider: number of processors and speed of each
processor.   Intel does not produce the fastest processors. (not by a
longshot!).  DEC alpha, SGI MIPS, Sparc, PA-RISC are all better at
number crunching.   The midrange to high-end computer market lept ahead
of the PC world years ago, although I think we will see things even out
in the next few years.

I was running on a Hewlett-Packard V2500 running 32 processors, each
running at 440MHz.  This processor (PA-RISC 8500 with 2MB onboard cache)
is capable of crunching a dataset in 4.5 hours.

Most of the people in the top 100 list are probably using UNIX-based
machines.  Some are using prototype hardware.   I'm intrigued by the
Intel performance group, myself!

Look at the number of hours that some have put in in relation to their
average time.  This is a good indication of how many machines they may
be running on.  Seti@sun must be running on 200-300 processors.

As far as performance on an intel chip, best bet is to run linux.  It
can cut dataset times down 20%.  In windows, be sure to disable the
graphics on the setiahome screensaver.  It wastes cycles.  You can
do this by setting the screen saver to go blank in 0 minutes.  (default
is to have the Seti graphics enabled)  I have seen data times go from 15 
hours down to about 10 hours just by disabling the windows graphics.  

Good luck,

-tjh

PS, I didn't complete all 40,000+ worksets on home equipment!  I am running
on a few machines at home now, though.