Pentium M's kick butt at...

[Jeff]

BOINC! I've been running one with Predictor@Home and SETI@Home for the past couple weeks and even at 2.17GHz and a lowly 135MHz FSB it is just kicking the crap out of my AXPs, DXs, and even my old A64. If you're into BOINC at all then I'd recommend these wholeheartidly(is that even a word ).

I'm tempted to pick up one of the setups that have been for sale recently here at XS because I couldn't even imagine what 2.6 or 2.7GHz at over 200MHz FSB would do!

[jjcom]

That'd be nice for me, very nice. What kind of numbers are you getting in SETI? From what I can tell, FSB doesn't seem to matter with SETI, at least for me. 166x13.5 does about the same as 220x10.5

m0p was trying to convert me to Pentium M...and this looks good...need to up my SETI output

jjcom

[Jeff]

Seems each SETI@Home unit is taking on average ~6400secs or 1.7hrs. And that's at 2.17GHz with 135MHz FSB. FSB should make a huge difference... or at least it did with the original S@H. I was very active in my S@H team's benchmark challenge and you could drop loads of time by upping the FSB. I had the lead until these P-Ms came out.

[jjcom]

6400seconds! I get about 9500seconds on my AXP. I might have to add somethings up and see how cheap I can get making a Pentium M rig...This is good news for me Jeff, thanks

[Mike@Portugal]

My P-M@2592Mhz does 3300secs of average in SETI@Home with the optimized client that can be found here:

http://www.marisan.nl/seti/index.htm

There are optimized clients for every cpu, p-m, p4, A64, etc.

[jjcom]

Too much money for me to get a Pentium M rig going and a A64 rig...so I'll just get a A64 rig going, that'll give me a nice boost to my stats

I'm going to do some comparison testing now...105x22 vs 220x10.5 on my Athlon XP. Should be interesting to see what the results will be

[Jeff]

Wow... thanks for the information Mike!

And I hear you about $$$ JJ... I'm really wanting to add another P-M but the $$$ is holding me back right now. However... I think I just sold a couple of pinball machines so I may have a little money to play with later this week.

[jjcom]

Looks like this WU's going to take me about 3 hours to do...my average is around 2 1/2 to 3 hours. Doesn't look like FSB makes too much of a difference, at least on Athlon XPs

jjcom

[Jeff]

Wow... I can't beleive BOINC is that much different than the old stand-alone version. Here's the benchmark table I referred to for SETI@Home classic...
http://www.users.globalnet.co.uk/~mu.../benchmark.htm

Note: My best result is over a year old and was just flying for that time.

[jjcom]

I was expecting at least a noticible difference, even though it doesn't seem to make that much difference, but I was surprised to see it making that little since it was such a huge FSB difference

[[XC] moddolicous]

Wait, is this the same as Folding@home, or just seti@home. BTW, Jeff & JJcom, can u run Dimes on your comps? There is a thread around here about that, just thought team XS could use your help.

[PS2pcGAMER]

Predictor is similar to FAH in that they both deal with proteins. Predictor is interesting in that it is participating in the CASP experiment which DF also did for a time. Predictor however uses the same platform as SETI (the BOINC platform) to distribute WU's.

Yeah, and I agree. Give DIMES a shot. XS is 13th in daily production already.

[Jeff]

Ok, I'm now "doing" Dimes(at least I was 'til the server went down).



And I just realized that in Predictor... the Pentium M is more inline with the A64... which is only a tiny bit better than AXPs. Code must not be optimized much. And another odd thing with Predictor is HT means NOTHING. Turning HT on basically just doubles the amount of time per processor so the output is exactly the same.