My New Hobby: Folding@home

[Ilander]

So, anybody over here at IW do folding@home? For those that don't know, it's a distributed computing project from Stanford wherein you can contribute processing cycles to protein-simulation.

Why do this? It helps aid researchers in cancer research, as well as Alzheimer's, Parkinson's, cystic fibrosis, mad cow disease...well, it's a long list of major diseases.

I got involved a couple of years back, when I first heard about it, but trying out on my Pentium 4 machine, couldn't really notice tangible results.

These days, though, in the era of multiple core processors, and unified-shader (DX10-capable) graphics cards, you can finish a work-unit in easily under a day, and noticeably help the cause...especially with high end gaming computers, as the graphics card clients are especially fast, thanks to having in some cases hundreds of separate processors to work with.

If anyone is interested, just let me know, I could even give you pretty good instructions on how to set it up on your computer.

Now, as for more depth, and how it's become my hobby? The real answer is that I'm a computer enthusiast, who loves hardware, and making computers run as fast as possible...and folding@home, ultimately, is a fantastic burn-in program, a good benchmark one, and, as a male, something to compete with, since Stanford University awards points based upon the amount of work you do for them.

They're not used for anything other than bragging rights, but I have to say...it's pretty cool to be able to say that I personally am in the top 1500 producers in the world...and soon to break the top 10,000 on the basis of total points.

There are stats pages where you can track your progress, like here, which contains my stats.

If anyone else on this forum is folding, please, share your EOC stats page.

[dplax]

I used SETI@home a couple years ago...

[Ilander]

Seti@home has similarities in execution of the standard client...and even the multi-core client...but, to my knowledge, Folding@home is the only distributed computing app that can really harness the power of a modern graphics card, which just has an order of magnitude more processing power overall...especially in scenarios that benefit from parallelism.

I guess I could include a link to it...

http://folding.stanford.edu/

[dplax]

Might have to check it out then. Afterall, I do have a pretty powerful machine...

[Cerek]

I do folding@home as well - mostly just towels and washcloths, though. Shirts get hung up till ironing, shorts go back in the drawer and bedsheets go back on the bed (cuz I really hate folding the bottom sheet).

I don't really track my stats, except to make sure nothing is leftover when I get done. And while I'm sufficient at the task, I doubt I'm one of the top global performers.

[Chewbacca]

I set my PS3 to do this a while back, but never followed up. Seems worthwhile, thanks for the reminder.

[Ilander]

I recommend using a box fan near your PS3...it adequately vents, but the fan helps keep that corner of the room from being a hot zone.

[ocellis]

I used to run it a few years ago, when i was still cooling my CPU with a homemade water cooling system. The increased electricity costs eventually began to bother me, and when i brought back a fan for cooling, the noise made me uninstall it.

[Ilander]

That's the story I've heard a lot recently...that many, many of us hobbyist folders have discovered it fairly early, been discouraged by the output relative to the heat, noise, and power...and moved on.

I'm here to tell you, though...that's not the case anymore. My stock Phenom II 940 runs at 51 degrees Celsius (same as my old P4) but generates close to 4000 points per day, which, basically, amounts to 3 complete simulations a day...pretty astounding, compared to the P4, which would generate one every five days or so.

A graphics card uses about the same power, and, depending, can generate up to 8000 points per day, or around 16 complete simulations.

Pretty amazing stuff considering that the proteins they're simulating have literally hundreds of atoms.

[ocellis]

Unfortunately, I'm not really concerned about the efficiency of the programme. All i care about is the noise and the electricity bill, both of which are invariant to the output.
Back when i contributed, there was no client for graphic cards with my chipset (geforce 7,i think). Even if there were one now, i'd still loathe to use it. The cpu fan noise is kept at a minimum due to a large passive cooling block and large rotation blades.
The graphic card is cooled by a fan with next to no passive element and tiny blades, which have to spin at a high frequency to be effective, causing a terrible racket.