F@H on Comenius project :)

[Daniel0108]

Hi!

I think / hope this is the right topic

I'm taking part in the Comenius project (http://en.wikipedia.org/wiki/Comenius_project).
The topic is "Cells". Especially cell folding I want to introduce Folding@Home in this project.
So, it would be nice if someone could tell me more about Folding@Home.
I already read the science page(http://folding.stanford.edu/English/Science). But I want to know if there is anything important, that is not on the science page (Or if the content of the science page is enough)
If my lecture of Folding@Home is good, maybe some schools will install Folding@Home on their computers and help curing diseases.
Some computers of my school are already taking part in the f@h project

Yours,
Daniel

[bruce]

I heard a presentation by the Comenius project at a recent conference. It is a good project.

The science described on the science page covers enough of the essential information that you'll be able to convince others to help with the project. There is a tremendous amount of important scientific information in the technical papers found here but they're Graduate Level research papers so it's unlikely that you'll be able to use any of it. There are a couple of videos in the papers or on youtube which may be used to demonstrate the complexity of the problems being solved. (I've used http://www.youtube.com/watch?v=gFcp2Xpd29I)

People should not expect that sort of graphics from their individual computers, though. You'll each be contributing a small segments of analysis that can be made into a composite like that which represents the collective work of many, many computers. (In comparison, what you'll see f@h doing is pretty boring and the graphics doesn't really work.)

Your individual contributions to this research are important, though, so thanks for your support of the project.

[Daniel0108]

Thanks for your answer!
That helped me much
I will watch such videos, so that I can explain protein folding
Some of the scientific information is also important, so I will search in these papers if I have some questions (Or I just ask here in the forum)

I am currently programming on a stats viewer for F@H So that the users can easily see how much they have folded. That's better than nothing

(In comparison, what you'll see f@h doing is pretty boring and the graphics doesn't really work.)

You mean, the graphics of the GPU client aren't real?! Are they just random models, or are they rough models of the real protein(because the real proteins are much complexer)?
One more question(a bit off-topic): Does the GPU client also work on Linux, and for other Graphic cards(not only nvidea)?

Yours,
Daniel

[bruce]

The GPU graphics is real but the software works poorly or crashes on many systems. It also takes a lot of computer time to display the graphics so it reduces the real scientific work.

I expect that Stanford will eventually develop a Linux GPU client but there isn't one now. There's a user-supported shell that works for Nvidia G80 GPUs inside of WINE.

A lot depends on who will be installing and running the clients. Do they want to install something and keep playing with it, tweaking things for increased productivity or dealing with bugs or do they want to set it and forget it. Will the machines run a few hours a day or 24x7? This information should NOT be in your initial presentation, though.

It's more important to get folks interested and avoid the complexity than to present too much information too soon. For school use, a stable client that will produce reliable results consistently is probably more important than getting high production along with the problems that are often associated with beta software. The uniprocessor client is extremely reliable but relatively slow. It can be installed as a service if people want to ignore it. The GPU2 client for NVidia g80 is also pretty robust and you do get graphics. The GPU3 client which supports the latest generation of NVidia (the 4xx series) is still evolving. The productivity of the GPU2 client for ATI is fairly low and the graphics probably doesn't work, but the client does. The SMP client is in good shape as long as you use the latest version.

Details like this really belong here: viewforum.php?f=61 instead of in the science section.

[Daniel0108]

Thanks for your information
That helped me much

One question left, tough
Now about 350.000 CPUs are folding, right? So... With this information, when do you think the project will finish?
How long does it take till we have enough information to cure the diseases like cancer, Alzheimer and so on?

I hope you can answer this question, because it is very interesting

PS:

Details like this really belong here: viewforum.php?f=61 instead of in the science section.

As I said, it's a bit off topic

Yours,
Daniel

[John Naylor]

Thanks for your information
That helped me much

One question left, tough
Now about 350.000 CPUs are folding, right? So... With this information, when do you think the project will finish?
How long does it take till we have enough information to cure the diseases like cancer, Alzheimer and so on?

This project alone will never solve the riddle of how to cure cancer or alzheimers or whatever. What this project does is fundamental research, which can then be used and developed by other scientists as they wish. It's more likely that Folding@home will provide an explanation for some basic problem, some other scientists will build on that, yet more scientists will build on that, some more will build still further, and then a pharmaceutical corporation (or even some professors working in a University just as those that control Folding@home do) will finally use all of this research to come up with some form of "cure". That however is probably some distance away, so the best answer we can probably give is "the project won't finish anytime soon"

(personally I'm betting another couple of decades at least )

[Daniel0108]

Okay, thanks

Then I ask you: when will the folding simulations finish?
So that the scientists can start building on that

the best answer we can probably give is "the project won't finish anytime soon"

Do you mean, it takes long time to simulate folding? Or do you mean, it takes long time to cure diseases?

Yours,
Daniel

[bruce]

Both.

Specific proteins take weeks to months to years to process, and for a complete understanding of a single protein, there's a lot that would take much longer than that, although computers are getting faster so there's still hope.

Folding@home is studying a number of proteins as well as interactions between specific areas on one protein with areas on another protein. Even if you don't understand the advanced scientific concepts, the list of scientific papers that I linked to earlier indicates the quantity of new things that have been learned by Folding@home.

Research isn't a linear process. Some new information leads to new questions. Other new information may help with several unsolved problems. It's difficult to predict exactly which new fact will lead to a "cure" All we can really say is that (1) Protein folding is a very complex field if knowledge and we're learning a lot about it, and (2) specific diseases are directly related to mis-folded proteins but we don't understand it well enough yet to be able to use what we know as a cure.

[Daniel0108]

Okay thanks

No questions left. I understood everything

Thanks to all posters

Yours,
Daniel