Stratcat,

On Thu, 05 Aug 2004 17:57:18 +0000, Stratcat wrote:

...

According to the 'tester', there was no difference. I don't know his
methodology. There's always the possiblitlity of a 'difference' in run
times being w/i the margin of his testing error.

Who's tested this?

I was reporting the results of some casual testing back when I got my
first Pentium 4-based system. Based on a handful of work units, both the
GUI and the CLI clients processed a typical work unit in around 4+ hours
(generally more than 240 minutes, generally less than 255 minutes).

Given the variability of computation demanded by each work unit (in the
current incarnation of SETI@home Classic), one would have to do much more
extensive, detailed and sophisticated testing to say more than "they're
about the same speed."

Now, let the GUI version show those silly grpahics and the per-WU time
goes way, way up. Based on that, there's no comparison. The CLI wins
hands-down.

Back in the days when I was using a Mac (400 MHz G3), I'd use AppleScript
to shut down SETI@home during backups and then to restart it and
especially make sure the graphics were hidden after restarting it when the
backup completed (backups happened while I was asleep).

My reply was in response to Randall's post, where he states he did a
back-to-back comparison. It can be found in its entirety here:

news:pan.2004.08.04.17.57.16.773924@tarpit.net

Lost in the Team Lamb Chop stats???

He did not claim any affiliation w/TLC in his post. That's not to say
he's not a TLC member. I do not recall him from the ARS forums, but it's
a very large team w/only a portion of the members active in the forums.

I don't even know what "Team Lamb Chop" is, but the name doesn't endear
me. Besides, I wouldn't join any group that would have me as a member.

This is now a little 'academic' as you should be moving over to BOINC.

True! ;-)

I've got BOINC installed and I run it when the servers can supply me with
work. As they're slowly getting the various wrinkles worked out and
are able to send me work units, I'm shuttind down Classic and using BOINC.
For one thing, I have a Hyper-Threading CPU and I find it easier to get
BOINC to use "both CPUs" than I do Classic.

I'm frankly a bit surprised at how much overlap you can get between to
concurrent SETI@home clients with Hyper-Threading. Basically, "top" shows
two full CPUs worth of utilization. Once things settle down, I'm going to
try to work out how the MFLOPS numbers compare betwen dual and single
SETI@home clients.

But it seems to have been a perennial question during the time I've
spent here...

Only because we're given a means of comparison, thus tweaking the
competitiveness parts of our psyches...


Randall Schulz