Folding @ Home and Distributed Folding

[Robor]

EDIT: Title changed to hopefully get more responses from respective team members...

Okay, what is the difference between Distributed Folding and Folding@Home? 😕

Are they the same? If not, what's the difference between them and are they related?

 

[Baldy18]

Originally posted by Muttley:

F@H deals with a specific set to work with.

DF approaches from a broad idea to isolate stuff that doesn't work and stuff that has a higher potential to work roughly 5 terra of memory or the size of a persons brain. Trials were going on using small and now larger proteins to target against using now 10 billion structures.

phase 1a:
1ENH: 100%, best RMSD: 3.64 (March 15, 2002)
1PMC: 100%, best RMSD: 3.96 (March 23, 2002)
5PTI: 100%, best RMSD: 5.40 (April 1, 2002)
1VII: 100%, best RMSD: 2.19 (April 8, 2002)
1SHG: 100%, best RMSD: 6.52 (April 15, 2002)

phase 1b:
4ICB: 34.0%, best RMSD: 4.61

CASP5 targets:
T0129: 100%, best pseudo-energy: 42.32 (June 11, 2002)
T0131: 100%, best pseudo-energy: 1.3 (June 18, 2002)
T0135: 100%, best pseudo-energy: 0.70 (June 25, 2002)
T0138: 100%, best pseudo-energy: 0.39 (July 3, 2002)
T0139: 100%, best pseudo-energy: 6.79 (July 9, 2002)
T0148: 100%, best pseudo-energy: 47.79 (July 16, 2002)
T0147N: 100%, best pseudo-energy: 22.35 (July 23, 2002)
T0157: 100%, best pseudo-energy: 54.42 (July 30, 2002)
T0156: 981.8 million, best pseudo-energy: 44.16

Proteins have a vast number of folds, larger than we could hope to compute even with distributed computing. Usually only one fold is found in nature. The Distributed Folding Project aims to test our new protein folding algorithm. We want see if it can reproduce natural protein folds after making extremely large samples of many different folds.

With your help, we will create the largest samples of protein folds ever computed. First we will make 1 Billion (1,000,000,000) folds for 5 small proteins, then we will try 10 Billion (10,000,000,000) for another 10 large proteins. By the end of our first phase, we hope to make over 100 Billion protein folds spanning 15 different proteins.

muttley

 

[Assimilator1]

I thought they were the same projects too😉 ,but their not ,from what little I know I believe they have similar goals but they are 2 seperate projects.
I'm sure 1 of the DF/F@H guys will speak up in amin

[edit]Baldy18 jumped in as I was writing!😉

 

[Robor]

(just a bump)

 

[Crazee]

Taken from the DF Site:

Proteins have a vast number of folds, larger than we could hope to compute even with distributed computing. Usually only one fold is found in nature. The Distributed Folding Project aims to test our new protein folding algorithm. We want see if it can reproduce natural protein folds after making extremely large samples of many different folds.

With your help, we will create the largest samples of protein folds ever computed. First we will make 1 Billion (1,000,000,000) folds for 5 small proteins, then we will try 10 Billion (10,000,000,000) for another 10 large proteins. By the end of our first phase, we hope to make over 100 Billion protein folds spanning 15 different proteins.

Taken from F@H Site:

What does Folding@Home do? Folding@Home is a distributed computing project which studies protein folding, misfolding, aggregation, and related diseases. We use novel computational methods and large scale distributed computing, to simulate timescales thousands to millions of times longer than previously acheived. This has allowed us to simulate folding for the first time, and to now direct our approach to examine folding related disease.

It looks to me that they have very similar goals. As to the client software and how they run I can't speak for F@H because I haven't run it yet, but I am running DF on a computer and it runs very smoothly. The stats at their site are updated every 10 minutes. It seems from what I have read that the client runs best on linux. P4s perform almost as well as Athlons clock for clock so if you have an overclocked P4 apparently this is a good project for you.

Hopefully someone who has run F@H can tell you more about the client and stats. I do know that TeAm DF is comprised of only 27 people and thus far only 24 have returned results.

I guess it wouldn't hurt to post some links 😉


Project Website
TeAm stats page
Statsman Distributed Folding Stats

We are currently #35 in the project and moving up. 🙂

 

[networkman]

Having now achieved Top 10 status on our Folding@Home team, it'd probably be a good idea for me to volunteer some information. 😉

When I started the F@H project, I wasn't aware of the DF project.. perhaps it hadn't even started yet, but it was time for me to get involved with another project that was different than what I'd done before. If you wish know more of my motives, PM me and I'll discuss it. 🙂

In any event.. let's see important stuff.. hmm, stats! 🙂

This LINK is to the stats of our own TeAm, #198, and is the official stats hosted at the Stanford site.

Statsman.org has a little bit more in the way of detailed stats, both for the teams and the individuals.

One thing you'll notice very quickly(especially so if you've done alot of SETI) is that not only is the number of completed WUs kept track of, but also the "score." Different work units may be assigned various weights depending on the time slice they represent, so it is quite possible to get WUs that are shorter or longer than others, with the shorter ones giving less of a score than the longer. Unfortunately, these are assigned at random - there is no way to look for the longer work units. 🙁 Your ranking on the ladders is based upon your total score, NOT the number of work units completed! This explains why someone with a greater number of completed work units can be farther down the ladder than someone with fewer. 😉

A major downside of the project is that there is no caching of work units available - sorry, no equivalent of SetiQ or SetiDriver for this project, at least not that I know of.. though I continue to hope for something. I rather suspect that as the project evolves, the staff in charge will realize the value of making additional improvements, just as other projects have.

I cannot give you any insight on the DF project as I've not gotten involved in it yet, but hopefully, this will help you out. 🙂

EDIT: Duh, I should make sure to post the link for F@H home page.. 😱

 

[Robor]

Okay, I'm HERE and want to install the client as a service under Win2K so I choose it and I see dasm!!!, an 8.19MB download for a CLI client? (downloading now) 😉

Client downloaded and I'm registered (I think). Waiting for the Email but it's late so I'm headed to bed...

 

[Crazee]

Once you have your handle, go to here and login in. You will have to enter the information for your profile and select the TeAm 🙂 The link I sent you in PM has more info on how to install it as a service if that is what you want to do. I run it with dfGUI at home, but I bet you want to run it as a service for testing purposes.

 

[Poritz]

I have been experiencing difficulty with the F@H client under XP Home so I am thinking about switching to DF.

-Poritz

 

[m0ti]

^ ^ ^
|| || ||

Up you go


(bump)

 

[networkman]

I don't have any clients at home or work running XP, thank goodness. 😉

 

[Crazee]

From everything I have heard the DF client runs very well in Win2K/XP. Where it seems to have some problems is running as a service in 98.

 

[Robor]

Okay... I guess I have the DF client installed. Questions...

How do I run it without a DOS window in my taskbar (would prefer as a hidden service)?

Is there any way to run it alongside SETI@Home? Run together SETI takes all the CPU.

Do you recommend letting it use 25MB or 100MB+? What's everyone else doing?

EDIT: Nuther question... I let it download the updated client. If I leave the update in the install directory will it have to redownload it if I install it on another machine?

Thanks!

 

[Crazee]

I use dfGUI to run the CLI with. It gives you a graphical interface to adjust the client with. I run it using the extra RAM since this machine has 512MB of RAM.

I have read that it is easy to setup as a service. If you are interested in that it is in the FAQ from the link I sent you via PM 🙂

 

[Robor]

Okay, here's what happened to me...

========================[ Aug 14, 2002 8:23 PM ]========================

========================[ Aug 14, 2002 8:25 PM ]========================

========================[ Aug 14, 2002 8:28 PM ]========================

========================[ Aug 14, 2002 8:48 PM ]========================

========================[ Aug 14, 2002 8:49 PM ]========================

========================[ Aug 14, 2002 8:52 PM ]========================
FATAL ERROR: [003.001] {foldtrajlite.c, line 4258} Unable to fetch Biostruc

Any ideas?

 

[Baldy18]

Originally posted by: Crazee
From everything I have heard the DF client runs very well in Win2K/XP. Where it seems to have some problems is running as a service in 98.

Not really a problem running it as a service in Windows 98 as much as being able to run it at all in Windows 98.:disgust:

Originally posted by: Robor
Okay... I guess I have the DF client installed. Questions...

How do I run it without a DOS window in my taskbar (would prefer as a hidden service)?

Is there any way to run it alongside SETI@Home? Run together SETI takes all the CPU.

Do you recommend letting it use 25MB or 100MB+? What's everyone else doing?

EDIT: Nuther question... I let it download the updated client. If I leave the update in the install directory will it have to redownload it if I install it on another machine?

Thanks!

If you are using Windows NT/2K/XP follow the service install directions found here. They are very simple and it works great.

DF priority can be changed between +20 and -20. I haven't tried it but you could probably find a priority that will alow you to run it alongside SETI.

If you have 256MB of RAM or more I suggest using the 100MB+ method since it will double your production and doesn't seem to hurt performance unless you start using the paging file a lot.

I believe that ever pc will have to download the update but I'm really not sure. Please let us know what the results are if you test this.

 

[Crazee]

Originally posted by: Robor
Okay, here's what happened to me...

========================[ Aug 14, 2002 8:23 PM ]========================

========================[ Aug 14, 2002 8:25 PM ]========================

========================[ Aug 14, 2002 8:28 PM ]========================

========================[ Aug 14, 2002 8:48 PM ]========================

========================[ Aug 14, 2002 8:49 PM ]========================

========================[ Aug 14, 2002 8:52 PM ]========================
FATAL ERROR: [003.001] {foldtrajlite.c, line 4258} Unable to fetch Biostruc

Any ideas?

From the known bugs and fixes:

'Unable to Fetch Biostruc' error message
This is not a bug. This occurs if you erase your TEMP (/tmp on UNIX) directory while the software is running. Do not erase files in your TEMP directory while programs are running. Some of these are actually being used by the software

 

[Robor]

Hmmm... I didn't delete any files while it was running. The only thing I did was try playing with the "nice" switches. I kept bumping the DF client up in priority. Started at 10 then 1 then -10. Either I was doing something wrong or they didn't have any impact. I did this while SETI was running to see if I could get the 2 proggies to share the CPU. Never happened. The one time I couldn't "ctrl-C" the DOS window and had to just close it and kill the process in task manager. Maybe that's the cause of the error? 😕

I'll look into the service as soon as I get a chance. Thanks!