Folding@Home proof?

Ease · 21 Jan 2009 at 18:11

So I was wondering, please take no offence, but is there any proof that F@H actually helps? I mean sure, you're processing thousand upon hundreds of gigabytes of data, but has there been any actual achievements from it?

It seems like there's a bit of effort to put in to get it all set-up, and don't want to bother with it unless I am sure that it will actually help.

Hatter The Mad · 21 Jan 2009 at 18:18

It will help in the same way that digitizing and sharing the human genome has helped - the information is there, tests and research and be done faster based on this data.

A.N.Other · 21 Jan 2009 at 18:31

http://folding.stanford.edu/English/Papers

muppetkiller · 21 Jan 2009 at 22:37

Every little helps.

Ease · 21 Jan 2009 at 22:39

Alright, I'm running three Consoles now. They'll be on about 20 hours a day, 7 days a week. It says I'm getting 220PPD? That doesn't sound like a lot...

vertica · 21 Jan 2009 at 22:48

I think you may be better off running just two clients.

1 SMP client on you quad should see ~1500-2000ppd if left 24/7
1 ATI GPU client on your graphics card should see another 2k ppd (the ATI client just isnt as good as the nVidia one)
still you should see 3.5k ppd average.

all depends on actual up-time, the work units you get, random errors, power-outages etc

the SMP client on your quad will probably take ~25-30hrs to do a single unit
the GPU client ~4hrs per unit (this is a guess as my experience is with NV cards)

Ease · 21 Jan 2009 at 23:05

I tried the ATI GPU, and my machine crashed for the first time in months (and last time it only crashed because I was playing with overclocking...) - so I've decided against using it. It might have been a fluke, but I have a lot of important projects at the moment and can't risk the inconvenience.

ChrissyT88 · 21 Jan 2009 at 23:31

I would definately use the SMP client if you are going to be running the machine for 20hrs a day. It should get you a lot more points (probably closer to 2000 or more depending on your setup), but the dealines are a lot shorter. However, 20 hrs a day should not be a problem!

I think F@H is quite scientifically relevent - essentially, the program just runs molecular dynamics calculations that would traditionally be run on a supercomputer or a large cluster. Obviously this is expensive, and the DC elecment of the project is a novel way to get around this limitation, but the science is well grounded. If we were not running the program, then chances are it would be running (a lot more slowly!) elsewhere. The research looks promising, but, as in all research, there is a chance that it may not lead anywhere. However, if me leaving my PC on for a bit longer helps one person, or opens up a new topics for research, it think it would be worth it (however corny that sounds!

).