How slow can you go?

I have a lot of old machines, and I want to get a few of them folding, as they aren't really being used for anything. Trouble is, they're slow. So I ask you, is a P2-300 fast enough for F@H? How about a P120? A P75? I'm not going to bother with the 486es.

I have a P166 at work that is folding using the Google Toolbar hack that takes about 2 weeks to do a WU but it doesn't run 24/7. The 350 will take a couple days running 24/7, the P120 and P75 may take a while.

they benchmark the WUs on a p500 celeron that runs 10hrs a day.

as long as you let them run like all the time you might make deadlines

Haw!
I have...

2 Pentium-Pro 200 folding 24/7
1 Dual-Pentium Pro 200 folding 24/7
4 P3-350's folding 24/7
2 P3-450's folding 24/7
6 P3-733's folding 24/7
all the rest (11 or so) are 1ghz or higher... I plan on becoming the undisputed king of old-school folding.

vonkaar said:

Haw!
I have...

2 Pentium-Pro 200 folding 24/7
1 Dual-Pentium Pro 200 folding 24/7
4 P3-350's folding 24/7
2 P3-450's folding 24/7
6 P3-733's folding 24/7
all the rest (11 or so) are 1ghz or higher... I plan on becoming the undisputed king of old-school folding.

Click to expand...

That must REALLY add to the electric bill!

Axle

I don't know... the company I work for hasn't ever mentioned anything =p. They've been in the bad habit of never logging off for years, so this should be nothing new. Plus, It's been going on for 2 months... in Texas heat... soooo.. no problem =p.

I have about 5.5ghz at home... but my fiance pays the electric bill (her only bill... grr...) and she hasn't complained... sooo... haw!

My recommendation for anything older than a Pentium II is to run Genome work units only. You'll still get credit for yourself and the team but there is no deadline on those work units.

-Rav

DoorBasher said:

I have a lot of old machines, and I want to get a few of them folding, as they aren't really being used for anything. Trouble is, they're slow. So I ask you, is a P2-300 fast enough for F@H? How about a P120? A P75? I'm not going to bother with the 486es.

Click to expand...

<sigh> I've seen a couple of posts like this recently.

Your p2-300 MIGHT be okay.

The underlying question is not whether a box will FINISH a WU - I'm sure even the slowest 486 will eventually finish grinding out a WU. The REAL question is a matter of timing.

To be of any benefit to the research, a F@H WU must be returned to Stanford within the deadline. If it is not, you still get credit for it, still earn points, etc. ... but the server re-assigns it to another folder. The net result is that completion of the WU was delayed by one deadline.

The standing suggestion for contributing slower CPU cycles to the cause is to run GENOME at Home WUs on them. The G@H project has no deadlines, and is related to the F@H project (don't ask me specifically how - I don't even pretend to understand) so the slow boxen still make a relavent difference.

Please understand - I am not attacking anyone here, or saying we don't want your boxen. But there is a proper application for different speed CPUs, and speed matters.

All of my 'old' machines finish the WU well within their deadlines. The closest I've ever come was a week or so til due-date.

If you run them 24/7, then you should be able to hand them in on time. Well, the P2-300, at least. The others would be best used doing genomes, but you could give the faster one a shot at the tinker core. Let it run for a bit, then use EMIII to monitor it, and see if it will finish in time.

Rav said:

My recommendation for anything older than a Pentium II is to run Genome work units only. You'll still get credit for yourself and the team but there is no deadline on those work units.

-Rav

Click to expand...

I was going to reccomend the same.

vonkaar said:

All of my 'old' machines finish the WU well within their deadlines. The closest I've ever come was a week or so til due-date.

Click to expand...

Weeks? Lat I looked most of my wu's had deadlines measured in mere days. I havent looked 4 awile tho.

Current Work Unit
-----------------
Name: p680_L939_K12M_414K
Download time: August 15 16:21:00
Due time: October 4 16:21:00
Progress: 68% [||||||____]

Then it's settled. The P2s will fold, and the rest will run genome. I might see if I can squeeze a few more MHz out of the P2s, but I doubt it, since they're already overclocked by 67 MHz.