The Definitive Folding @ Home Thread - FAQ, Links, Info and More!

[DSwarP]

The Definitive Folding @ Home Thread

This thread is a compilation of previous threads made by Jason, HotRod, and Hack99. Per Jason's request, I am trying to update and consolidate all the Folding @ Home information into this thread. If you have suggestions, please let me know.

Taken from the Stanford University Folding @ Home Website

Our goal: to understand protein folding, misfolding, and related diseases

What is protein folding and how is folding linked to disease?
Proteins are biology's workhorses -- its "nanomachines." Before proteins can carry out these important functions, they assemble themselves, or "fold." The process of protein folding, while critical and fundamental to virtually all of biology, in many ways remains a mystery.

Moreover, when proteins do not fold correctly (i.e. "misfold"), there can be serious consequences, including many well known diseases, such as Alzheimer's, Mad Cow (BSE), CJD, ALS, Huntington's, Parkinson's disease, and many Cancers and cancer-related syndromes.

You can help by simply running a piece of software.
Folding@Home is a distributed computing project -- people from through out the world download and run software to band together to make one of the largest supercomputers in the world. Every computer makes the project closer to our goals.

Folding@Home uses novel computational methods coupled to distributed computing, to simulate problems thousands to millions of times more challenging than previously achieved.

What have we done so far?
We have had several successes. You can read about them on our Science page, Results section, or go directly to our press and papers page.

Want to learn more?
You can download our Executive Summary, which is a PDF suitable for distribution. Also, you can learn more by watching recent seminars (Stanford BMI ; Xerox PARC). One can also help by donating funds to the project, via Stanford University.

-------------------------------------------------------------------------------------

Quick Links

Basic Info About F@H:
Stanford Folding@Home Webpage - The official site for getting the F@H client and other various resources.
Official F@H Forum - This is the Official Stanford F@H forum.
Tweak Guide By [Ch]amsalot - Good Info on setting up F@H optimially.

Specific Technical Problems:
Freezing with AMD and SSE? - The official thread if you are experiencing your AMD system locking up with the -forcesse tag.
How to Fix "MISSING_WORK_FILES" Error - A good guide for fixing Missing Work Files.....Thanks to FemFolder & Sombodynew

Stats & Related:
EOC F@H Stats Page - The BEST source for your folding stats.
Stanford's Stats Pages - Realtime, but a little slow, and doesn't give historical information.
- F@H Benchmarking ORB
- XTremeCPU FAH Team List


Folding Farms & Diskless Resources:
Overclockix. - Based on Knoppix but tailored for folding.
GAHnix - A small compact Linux distribution based on Trinux, for disk-less network booting nodes.
Building A Diskless Folding @ Home Farm - Article on setting up a diskless folding cluster.
Making ATX "Y" Cable - Power two boards w/one PSU. Alternatively, you can just purchase one here.

Monitoring Utilities:
FAHStats - Folding@Home monitoring program
HFM.NET - Folding@Home client monitoring application
Fahmon - Another monitoring program
FahSpy - Another monitoring program

General Utilities:
How to install VMWare Server under Windows for SMP folding for ver. 1.0.6 or later - Killerbeagle's Procedure
How to install Ubuntu(64 bit) and SMP client - cuog's Procedure
Simple how to: vmware player and NotFred's folding appliance - MtM Procedure
Tray-It - A utility which allows you to autostart F@H and minimize it to the system tray.
StressCPU2 - Utility designed to stress SSE on AMD CPUs. Based on a tweaked gromacs core I believe. Good utility for Overclockers to test stability!
CPUBurn - Another CPU Stress Utility
Graphical Signature Designer Program - Excellent program to add colors to your folding sig Forum thread with more info

-------------------------------------------------------------------------------------

FAQ

What is this “Folding at Home” stuff anyway?
Folding@Home (abbreviated as FAH) is a distributed computing project by Stanford University that studies protein folding, misfolding, aggregation, and related diseases. In short, they’re trying to find cures for diseases like Alzheimer’s, Mad Cow, Parkinson’s, CJD, and ALS.

Sounds great, but I have one VERY important question… won’t running this program hog my resources and bog down my computer?
The answer is NO! FAH only runs on unused clock cycles, and gives up ones in use as soon as another program needs them. A very good look into the performance impact of folding can be seen here.

Awesome, how do I get started?
All you need is a computer and an Internet connection. Just go to the STANFORD SITE, download the appropriate client and install it on your computer. Installation varies by client type and is covered below.

What platforms are the clients available for?
There are currently clients for Windows, Linux, MacOS X, GPU, and PS3.

What are the differences between the clients? For PCs, there are graphical clients, console clients, and GPU clients. For Macs, the options are the same, except no GPU client is offered. For Linux, there are Console versions only.

The Consoles run in the background and are launched from a terminal, to which they output text messages about the simulation's progress. The Graphical client runs in the background and has a graphics window available that allows you to see the protein being simulated.

The GPU and SMP clients are special clients. The GPU console client is designed for folding on a graphics card. The SMP console client is designed for folding on multi-core processors.

Which client should I use?
The graphical client is the easiest to install and to manage. It has a graphics window available that allows you to see the protein being simulated. It is installed via a simple installation wizard that asks you all relevant configuration questions. During installation, it will create a shortcut to the exe and place it in your startup folder so that it launches when you login. It will also put an icon in your system tray so you can minimize, maximize or configure the client. On the down side, the graphical client is known to interfere with some games.

The text only console provides information about the simulation in a text format. It runs in a terminal window and minimizes to your taskbar. You can use an app like Trayit to minimize to your system tray. You will need to create a shortcut manually and install it in your startup folder. The text console can also be installed as a service so it launches with Windows and will run in the background regardless of which user is logged in.

Which client is fastest?
Text > Graphical

How do I install my client?
**Make sure you enter 11314 as the team number**
**You can use any name you like but you can’t transfer points later if you want to change it**

To avoid confusion, you want a unique user name. The download page at Stanford has a name search window to help with name selection.

The download page can be found here: http://folding.stanford.edu/download.html

Installation instructions for each specific client can be found at the link next to each one, under the “Installation Instructions” column.

What are flags?
Flags are special instructions given to the client at launch.

You can see the complete list HERE

How do I use flags?
Append them to the target line in the shortcut that launches your folding client.

See the attached screenshot for an example.

Which flags should I use?
AMD Single processor rig: -advmethods -verbosity 9
Intel Single processor rig: -advmethods -forceasm -verbosity 9
AMD Multiprocessor rig: -local -advmethods -verbosity 9
Intel MP/HT rig: -local -advmethods -forceasm -verbosity 9

-advmethods: "Request to be assigned any new Cores or work units." In effect, you get gromacs most of the time. You will still get the occasional Tinker.
In April of 2004, the points awarded for many WUs, including tinkers, were increased. As a result, there is better ppd parity between tinkers and gromacs and the -advmethods flag now provides less benefit than it once did. In fact, some folders no longer use it.

-forceasm: "Instruct the core to use SSE/3DNow assembly instructions if possible, even if it has previously made the determination that the machine may not be handling this well."
forceasm is only effective if you have issues with the PC folding. By default a RIG (AMD or INTEL) will use SSE/3DNOW ECT... with no flags... but if you have a error or a reboot caused by instability in the PC or heat or for whatever reason... the FORCE XXX flag will force the PC to use SSE even after a crash ect... without the flag it will revert back to not using SSE to gain stability on a crash."

-verbosity 9: Sets the level of output detail that is written to the log file. V-verbosity 4, for example, will output less than -verbosity 9.

-oneunit: Stops folding after the current work unit completes. Useful if you want/need to do a format and install or stop folding on that machine. If you don't do this, you will probably end up downloading a new WU and then deleting it.

-config: If you want to change any settings, like priority, user name, team number, etc., it will run the configuration setup again., and continue the program after you complete all steps. The -configonly flag will only run the config setup, and then stop.

What core priority should I use?
Use low priority on a dedicated folding rig. The core will try to run at all times.
Use idle priority on an everyday use rig. The core will only try to run when NOTHING else is active.

Can I adjust the core priority using task manager?
No. The priority has to be set during configuration.

What do I do if I have more than one rig folding?
You set up each machine the same way. Use the same user name and team number no matter how many rigs you have.

Do I need a different machine ID for each rig?
No. The machine ID referrs to the number of processors in a single rig, not the number of rigs in your farm. Each instance of FAH needs a different machine ID, but each rig can have the same machine ID.

How do I set up a multi-processor/HT machine?
**You will need to use the text console client**

For true multi-processor machines, the new SMP client should be used. Follow the install instructions from the download page. As of right now, the “Install as a service” mode does not work on the SMP client. You can, however, create a shortcut to the EXE and place it in the startup folder, and it should work.

For machines with HyperThreading, the use of the SMP client is nor recommended. The newest text console is your safest bet.

What should I know about my FAH directory?
Work Folder – Contains the files your FAH client is currently working on or waiting to upload.
client.cfg – Contains the configuration information for your client. You should not edit this with a text editor, rather you should always use the -config or -configonly flag to make modifications.
FAHlog.txt – This is the log file of your clients activity. The “-verbosity 9” flag adds details to the log.
unitinfo.txt – This gives you the progress your client has made on the current WU.
queue.dat – This dat file is the que of WUs waiting to be turned in plus info on your current WU

What is a WU?
A WU is a work unit. It’s the term Stanford uses for a single simulation run.

Are all WUs the same?
No. There are three types of WU:

- FAHCore_65.exe - Tinker - 100, 200 or 400 frames. They are generally the slowest WUs and the -advmethods flag reduces the number your rig will be assigned unless Stanford releases a new crop.

- FAHCore_78.exe - Single Gromacs - 100 frames. Use SSE. They are generally the fastest WUs.

- FAHCore_79.exe - Double Gromacs - 100 frames. Use SSE2 They are much faster than single Gromacs on P4 rigs.

- FAHCore_A1.exe – SMP Gromacs - 100 frames.

There are many different proteins of each type being studied. See the list HERE

How long does it take to finish a WU?
A WU can take anywhere from 4 to 36 hours to complete on a fast rig. Obviously longer on slower rigs.
Stanford has set the scoring system so that you get more points for the bigger WUs.

How does the scoring system work?
This is how Stanford explains it:

“How do you determine how many points a work unit is worth? Before putting out any new work unit, we benchmark it on a dedicated 2.8GHz P4 with SSE2 optimizations disabled for the test. We plug the results of this into the following formula:
points = 3.5 * multiplier * (daysPerWU)
where daysPerWU is -- no surprise -- the number of days it took to complete the unit.”

My current WU is corrupt, How do I delete It?
- Close FAH
- Delete the Work folder and queue.dat file in your FAH directory
- You might also need to delete the FahCore_XX.exe file(s) so the client can re-download that too.
- Restart FAH

My log file says I can’t send/receive WUs, What do I do?
If everything is working on your end, check your log file for the IP address of the server your client is trying to connect to. Then check HERE for the server status. It may be a problem at Stanford.

How will my computer perform?
A good source of information about PPD (points per day) on specific rigs can be found at www.fahinfo.org

http://www.fahinfo.org/index.php?all...=true&offset=0
http://www.fahinfo.org/complete_wus.html.gz

GPU Performance Information
Excel file of nVidia Performance Statistics
Excel file of ATI Performance Statistics

How much bandwidth will FAH use?
You can expect FAH to transmit 2Mb ~ 3Mb in the process of uploading a finishsed WU and downloading a new one.

My WU won't upload, what is wrong?
There is a possibility that the upload server you are trying to connect to is down. Check can check your server status here.

How do I check my score?
You can check the official Stanford stats HERE

You can check the EOC team stats HERE

Where did all my WUs go?
It sometimes takes a few hours to a day or so for WUs to make it into the stats.
Double check your user name and team number and just be patient.

How do I monitor multiple rigs?
There are several applications that let you track multiple rigs from one machine:

HFM.NET
Client Monitoring Application for the Folding@Home Distributed Computing Project
Forum thread with more info

FAHSTATS
Basic Functions:
Track multiple rigs from network and view on local machine or on the Internet via it's built in Http server - Thanks (907)FAN-BOY
Estimates ppd, ppw for current WU
Maintains a list of Stanford projects
Changeable skins

DCMonitor
Basic Functions:
Track multiple rigs via network
Reports WU data as HTML

How do I get that Liquid Ninja sig thingy?
Add the following to your sig for the liquid ninja stats:

[img]http://www.liquidninjas.net/dc/ln_dc_sigs.php?xfahuser=YOURUSERNAMEHERE&xfahteam= 11314[/img*]

Remember to remove the * and add your username

What’s wrong with my Liquid Ninja sig thingy?
The liquid ninja stats run a day or two behind. Double check your sig and just be patient.

How do I get that other sig thingy?
Add the following to your sig for the statgfx stats:

[img]http://gfx.statgfx.com/old/folding.cgi?&username=YOUR_USERNAME_HERE&teamid=11 314&border=0,0,0&custom=0,0,255&label=0,0,255&head er=0,0,255&stats=0,0,255&bgcolor=0,0,0&trans=yes&t emplate=fah_original&.jpg&cpu=YOUR_CPU_COUNT_HERE[/img*]

Remember to remove the * and add your username and cpu count
You can change the colors if you like.

-------------------------------------------------------------------------------------

More Information

DanEnsign, who is a part of the Panda Group at Stanford, wrote a good article on "What is folding?".
It's a little technical, but it really explains proteins and what we are doing.
To view the article, click the show spoiler button.

DanEnsign:

In the links Ivoshiee gave, there are good comments from Prof. Pande about what run, clone, and generation are, but they're short comments. (Necessarily so -- the guy is more Boss than Bruce Springstein, and about that busy.) Hopefully the following is not only more thorough but enlightening. And not too long (yeah, right).

First, let's review some basic physics. The key idea is that of a "trajectory." You might recall Newton's Second Law, F = ma, which means that the acceleration a (change in velocity) that a particle experiences is proportional (by its mass m) to the force F it experiences. This means that if we can catalog all the forces on a particle, we can determine its acceleration. If we know the acceleration, then we can use calculus to determine the particle's position as a function of time, for all time. The result is what's called a 'trajectory' -- a kind of map of where the particle has been and where it will be going. By the way, when I say 'particle,' I mean that we could perform this analysis on atoms, protein molecules, baseballs, the space shuttle, the Sun, or anything in between.

The analysis gets a lot harder the more particles there are in the system -- for instance, if you set up a system with the Earth and the Sun as two particles, experiencing each others' gravity, then you can solve Newton's Second Law very easily and write down a function which describes the position of the Sun and the Earth at all times. If you include the moon or other planets, then you can't write down functions like this, though you can solve Newton II numerically. This is what we do for FAH -- solve Newton II numerically for thousands of atoms, thousands of times, once every femtosecond or so (that's "ten-to-the-minus-15" seconds). What we get is a trajectory for the protein atoms.

If we're simulating protein folding, then perhaps the trajectory will result in a folded protein. Perhaps not -- we don't have a way to say for sure how this happens for an arbitrary starting conformation. (But we're studying it, obviously, thanks to our Army of Undea -- oops, I mean FAH clients. The Army of Undead is for a different project entirely.)

Now, on my desktop machine at work, I can simulate a system of about 16,000 atoms moving for 1 nanosecond (ns, or "ten-to-the-minus-9" seconds) in one day. But the protein that I'm folding requires (on average) one microsecond ("ten-to-the-minus-6" seconds) to fold -- and this is a system engineered to fold fast. To get to one microsecond on my desktop machine, I'd have to fold for 1,000 days. Forget about "average" proteins, which might take hundreds of microseconds, or milliseconds, to fold.

Maybe I'd get lucky and the protein would fold in that time; maybe I wouldn't, and they'd find me 35 years later, in some sub-subbasement below the chemistry building at Stanford, a raving lunatic lost to the dredges of Ph. D. research, sneaking out only at night to feed on spilled yeast extract and collecting discarded NMR tubes to wear as primitive jewelry. (I heard this happened to a guy.)

To avoid life-wasting tragedy, we (and when I say "we" I mean, "Someone besides me, but who I know") has recruited hundreds of thousands of generous and interested persons ("you guys") to give us a hand with some of this work. I could run a trajectory for 1,000 days, but instead we've taken a shortcut and decided to run 1,000 or 10,0000 or 100,000 trajectories for a few days (or months or years) instead. On average, a few of these trajectories will result in a folded protein (and we have ways of yielding interesting and important information from all of the work done on FAH).

Okay, here it is: The CLONE numbers are labels for each trajectory that we run. Each GENeration is another chunk of time along that trajectory. So, say that I benchmark CLONE0, GEN0 (the first 4 ns). That WU is then done, and the FAH software builds a new WU with starting coordinates (and velocities and stuff) where mine left off. Then the new WU -- GEN1 of CLONE0 -- gets sent to you, and you simulate the next 4 ns. And so on. So CLONE is a label for an individual trajectory, and GENerations are time steps along that trajectory.

RUNs are groups of similar CLONEs. All the CLONEs in a RUN have the exact same atoms, the exact atom positions, the same temperature, etc. The difference is the starting velocities -- the initial motions of all the atoms in the protein are randomized. Although statistically the velocities are determined by the temperature, there are countless ways of partitioning the velocities to the atoms, so we try out 100 or so CLONEs to get a good feel for the sample space. Assigning different velocity sets to the atoms turns out to be wildly important: if the conformation we start with happens to represent the transition state (sort of halfway from folded, halfway from unfolded) then 50 of our 100 CLONEs will fold, and 50 won't.

The different RUNs in a PROJect might, in their simplest form, represent different starting conformations. So, we could start off 100 RUNs of different partially unfolded structures and try to find the one for which half of its CLONEs fold -- then that RUN has the conformation of a representative of the transition state.

So why is this transition state doohickey so important? The folded state is relatively easy to identify, especially if experimentalists have determined the structure for the protein under scrutiny, or for a very similar one. The "unfolded state" is a bit harder, but we can generate unfolded conformations by, say, simulating the folded protein at high temperatures so it "melts," or we can thread the amino acid sequence on a set of randomly coiled noodles, or whatever. But the path which connects "unfolded" protein with folded protein is not so easy to get to -- but if we identify the transition state, then we've found (at least one of) the paths by which proteins fold, and that's research in protein folding.

The RUNs might also represent slightly different proteins -- for instance, different mutants of some protein. They might represent other things that I haven't thought of, but whatever they are they are similar enough to other RUNs in the same PROJect, that, well, they're part of the same project.

So to summarize, when I'm setting up a project, I might do the following:
1. Pick 100 different unfolded or partially unfolded conformations of my protein of interest. These become my RUNs.
2. Then, I set up 100 different CLONEs for each RUN. (Well, I don't actually set them up myself, I just run a program. But I run it really well. And intelligently. And I look good doing it.) Each CLONE contains one WU at this point.
3. Then, I let the (100 RUNs) x (100 CLONEs) = 10,000 WUs loose on the world ("you guys").
4. Then, I go have lunch.
5. I come back weeks later to find WUs crunched and GENerations progressing -- each of the original 10,000 WUs was the beginning of one trajectory, so at the end, I have 10,000 trajectories of 50 or 100 or more ns.
6. Finally, I sift through the data and learn something new about protein folding!

And so it goes. I'm still new at this, so I haven't actually done steps 4, 5, or 6 yet, but I've got a good handle on 1, 2, and 3, and now it's a matter of waiting (and doing 1, 2, and 3 a lot more).

Hope that helps! Now to find something else to distract me from doing homework . . .
Dan

Here is the thread
http://forum.folding-community.org/v...eb8975c#123997

[Fourth]

good. sticky.

though i would suggest explaining teams(why should i select 11314?)and what brigades are.

[DSwarP]

Will do. Also, adding all those links now...

[AkXb70]

IIRC, Stanford no longer uses the 500MHz celeron, but a 2.8GHz P4 with SSE2 optimizations disabled for the test.

[frozndevl]

We may not want to reference UDMon in the basic FAQ since Stanford doesn't want us to queue up WUs.

[DSwarP]

Check, check...

[frozndevl]

In the how will my computer perform, pointing to http://www.fahinfo.org/ as a good resource to see how processors perform

[DSwarP]

Originally Posted by frozndevl

In the how will my computer perform, pointing to http://www.fahinfo.org/ as a good resource to see how processors perform

I was thinking about doing that. I was looking for a list of numbers, but couldn't find any.

[chiefish]

I have them bookmarked at home , tonite when i get home ill link the ppd averages of cpus and the ppd per w/u, straight from stanfords pages.

[DSwarP]

Originally Posted by chiefish

I have them bookmarked at home , tonite when i get home ill link the ppd averages of cpus and the ppd per w/u, straight from stanfords pages.

Awesome. I might not really "work" on it anymore today (aside from updates people suggest) but I'll get to it this weekend/next week.

How is the format by the way? Easy to look through/navigate/understand?

[BigO]

Looks pretty good to me Dswarp, good job...

[chiefish]

yup looks like good formatting to me , ill post the links and then you may cut and paste them whenever you get to it.
heres a good link for ppd on different CPUs http://www.fahinfo.org/complete_wus.html.gz Should probly go under the How will my computer perform question above
And this is a good link to check the stanford servers status http://fah-web.stanford.edu/serverstat.html
Also for the "which client is fastest" question above it should be SMP, PS3 or gpu not positive which produces better ppd, then text client and lastly GUi client
Also for the question above "are all wus the same" we should include the gpu and smp units that are out now as well
Other than those i think it looks great.
Also i just saw this posted in another thread , instead of having to make your own splitter for the atx cables they sell them pre made here http://www.stayonline.com/detail.aspx?ID=7898 and for 7 bucks you cant really beat it i dont think.

[HotRod]

Good work guys. I see a bunch of brolen links that I will try to fix. If they are dead, I will just remove them.

[DSwarP]

I just went through the "Quick Links" section and corrected and deleted missing or outdated links.

[NM_JJ]

Great post! I'm glad someone took this on. I may be adding random comments as I look throught this instead of gathering all comments and putting them in at one time.

FAHinfo.org is not very intuitive when it comes to figuring out what ppd you will get. Need to mention the different outputs it gives like ppd/core/GHz or ppd/core depending on where you get the number. Also mention that some of the numbers could be wrong since they rely on people putting in their own numbers without any checking. Most numbers are entered right but one wrong number can pull the average off.

Additional Comment:

Has anyone looked into optimizing an overclock for folding? I've tried finding something but haven't found it. For example: is it better to go for low timings versus higher bandwidth? I may look into this but it will be specifically for my system. Are there just too many variables to even attempt an answer for this?

[ustorfixx]

Originally Posted by NM_JJ

Has anyone looked into optimizing an overclock for folding? I've tried finding something but haven't found it. For example: is it better to go for low timings versus higher bandwidth? I may look into this but it will be specifically for my system. Are there just too many variables to even attempt an answer for this?

It all depends on YOUR system.
In your sig -
Gigabyte DS3 rev 3.3 (G)MCH +0.2V; e6600 400x9, 1.45V, 52C folding; Crucial Tracers DDR2 1066 2x1GB 800MHz, 2.1V, 4-4-4-12

This system I doubt your gonna get much better with the anything as far as timings ect. I've played with both of my c2d rigs and honestly it didn't make any difference with any higher bandwidth than @ 800. Because its a trade off - 1066 then you lose your timings. The only thing you could try that would do anything is try to tighten your timings but you'll realize that its alot of work for nearly no benefit. Your already at 3.6ghz...SMP it and thats about all you can do. Altho I will say I think mine does better in VMware/Ubuntu than in WinSMP, just because WIN took alot more attention. If I do anything to either of these rigs I'm gonna make them dual boot with Ubuntu rather than VMware.

[NM_JJ]

Ustorfixx: Thanks for the advice. My OC was a shot in the dark and I haven't tried to optimize it yet. I will try to see how much higher I can take the CPU and then see if I can lower the volts on the memory and (G)MCH. I'm not sure if I want to try 3-3-3-? timings. I might also try to see how high I can go with the 4-4-4-12 timings.

[MrObvious]

Sounds like the more MHz, the faster you go . I think with Intels, the faster CPU makes a difference.

[chiefish]

Hey i am wondering whythis has never made it to the stickies to replace some of the older stuff we have up there yet?

[Erecshyrinol]

Hey, hey, hey! *dances like a madman*

New guy to the team, have posted in a couple of the threads already. Thanks for combining all of the useful stuff here, it is great for a newb like me!

Me = JosephDeathTyrantCromlechErecshyrinol

I doubt my stats will show up, as I am still working on my first WU (250 frames).

p2421_Ribo_ile_aux - GROMACS core