Folding@home is a massively parallel, volunteer-powered project in biomedical research -- as apparently much of the gaming world knows, but I was very hazy about it until Fentonparkninja clued me in. Rather than detail it here, I'll link these descriptions for those who are also new to it:

• Wikipedia page about Folding@home
• Front page, http://folding.stanford.edu

The reason for this thread is that I'd like feedback on the idea of forming a team amongst Simmers.

From a technical standpoint, all that's involved in "being on a team" is that you enter the team number in your Folding client interface, and then the contributions of your machine (crunching protein simulations when your Sims are sleeping) gets aggregated with others to form a team score. Eventually we see our aggregated score (and our rank against other teams) climbing, leading us to celebrate the glory of the Sim nation , as well as doing some practical service towards solving some very serious diseases.

Some particular questions for the forum --

1. Do we know of any such teams already up, that would make more sense to join instead of spawning another?
2. If a new team were formed in or near MTS (this being my main social life, no lie), do we suppose it would get at least a half-dozen participants? (Simmers who either start folding, or had already been doing it but without a team.)
3. Or am I overthinking this, and should just go for it? (The interface is easy enough.) In that case most likely I would give it an in-character name, as if the team were started at the Twallan Sunset Valley National Laboratory, to the greater glory of the Master Controller as well as the Sims.

Thanks in adv.,
Ronnie.

EDIT: If you're new to this thread, the decision was "go for it", and the in-world team page is:
http://www.simnationallabs.org/rese...-collaboration/

I haven't participated in it before, but I'd be happy to join the team.

Bless you RoseCity! I've decided to go for it, and have put up an explanatory page on an in-character site. Our team number is 226071.

Okay, I downloaded - I left the Passkey field blank, hope that was all right.

I'm signed in, and I asked for a passkey, cuz it said something about extra points?

I'll see if I can figure out what that "Passkey" mishegoss is about, and post something here if I get anywhere. I did use it once when I downloaded the client, but I have forgotten how and why.

EDIT: Here is a page that explains "passkey" better than I can.

Some high points:

• Setting one is optional.
• Its main purpose is to distinguish you from others that might have or spoof your username, to ensure that points are tallied correctly (especially if someone else uses your name and tries to game the system.)
• It only comes into play once and then your client memorizes it. The only reason you might need to remember it, apparently, is so you can set up clients on multiple host machines, to inform each of those clients (and thereby the server) that they should be credited to you.
• Your score will probably rise a little faster if you set one, because there is a bonus for multi-core machines that turn jobs around reliably, and that bonus only goes to users who set a passkey.

Just a note: There are settings available if you right-click its icon (in the tray). My computer wasn't doing so well with medium (75% cpu). My fan started running all the time, so I backed it down to light (50% cpu), and it's been great. I'm guessing some of us might want to throttle it down even further while we're in SIMS. It's nice that they've got something to allow us to adjust according to our needs.

I'll experiment tonight, and see if it interferes. If so, I'll throttle back during game time, and then bring it back, otherwise.

It's been on light for 50 minutes now, and the fan hasn't had to come on once. And I can certainly say that it isn't interfering with any of my other processes. I've got 4 windows open on the net; (3 explorer and 1 firefox) I've got Word open to two documents; I'm in Paintshop Pro; and I'm playing music from youtube...all without lags.

Well, it's been a day now, and for some reason I have no points. The 5 WU I contributed are showing up on our team score, but apparently I haven't produced any "points" for us. I'm not sure if this is a function of having obtained that key...or if it's something to do with the large WU number?

I've been running my stuff back and forth from "light" to "medium" so...50-75% cpu.

I see that, and I don't get what it means either. Five at one whack is a lot of WU -- usually one takes far longer to process than to upload. Do you think those were five real ones, or could they have been five attempts to deliver one? I wonder if something got fouled at upload-time that kept it from being credited properly, or maybe triggered the cheat-detector. (This is a wild guess, based on Fenton's telling me he'd had trouble with an upload, of course after investing hours and hours to get the WU done.)

Do you have one running now that gives you an estimate of the time until your next upload? I'll try to keep an eye out for it.

I assume it's running. My taskmaster says I'm using 75% cpu...but I have no idea how to tell if it has an estimated time of completion. Is there a page for that?

NEVERMIND. I just found the page by clicking on the icon. ....it says I have nothing running. Nothing is getting folded. WTF. Do I need to uninstall and reinstall?

Ah. I found a big page with a lit planet earth at night. It says, "protein loading." Apparently I had failure. "one or more folding slot failed."
Something bothers me, though...it has no name listed. Donor: Name: Team: (all blank)

And now I've finally found the page I was on yesterday. Gotta be smarter than the mouse, right?



about 1100 points per day...that means I should have about 1100 points...unless it's awarded after completion only. Let's hope it doesn't fail again. I hope it's not my computer messing things up.

I'm reading the log, and in SEVERAL lines it reads:

An existing connection was forcibly closed by the remote host.

Is "the remote host" my ISP? Cuz my computer has been on non-stop.

Anyway, there are probably 20 of these in my log. The last one was just an hour ago.

Update - 1627hrs MST: I have posted a question at their newbie forum, asking about this mysterious "remote host," and asked them how I can stop him/her from forcibly closing my connection.

It's in moderation right now...so I have no idea when they will be answering the question. Maybe by tomorrow I will be pulling my fair share.

Update - 2033hrs MST: Sure enough, my computer wasn't doing well with something they sent, and it screwed 5 times. That's why my thing says 5 WU. None of them took. They say it's a common problem with people who have my graphics card...apparently it's a pretty shitty one. But I've been told that before.

I can tell you that whatever project got loaded this evening is still running. It's 30% completed. I was told there are jobs that do not complete in a day, and apparently one is not awarded credit in increments. Apparently they are awarded when the entire project is completed? They said the one I'm on will not be completed for another 3 or 4 days. WOW. So my score will be zero, and then alluva sudden it won't. Estimated points at the end of these five days will be around 6K. (Looking at this, I think the project will be completed in 2 more days, actually.



So until I get a decent graphics card, I won't be contributing big numbers. The card I will be buying is an AMD Radeon 6670. And I guess this cinches it, I will be picking it up with my next VA check. So the beginning of March, I will take my pooter to best buy, and have them plug me in a new card.


eta: I will take my pooter to best buy, and have them plug me in a new card. Is it just me, or did that sound vaguely dirty?

Tsyokawe, you're a national hero. I'll be very interested to hear what you find out. Since the project is huge, presumably nearly anything that could happen with any of us has already happened to a lot of people, whereby we can hope that they'll have an answer.

I would bet that the "slot failed" means that it tried to run on your GPU and couldn't work it out. That would not be a big worry; the main question is still what happened to the jobs that apparently did complete on the other slot (the main CPU) and then bolluxed up in trying to get back to the server.

lol - I love the quote on our team page from Dr. Susan Wainwright.

Early today, the folding team of the Twallan Sunset Valley National Laboratory exceeded 300k total points and overtook team #225029 representing the Aperture Science Enrichment Center.

Aperture Laboratories is a major U.S. supplier of shower curtains and classified technology, with offices in Ohio and large R&D facilities in Michigan. Its Folding team was established in August 2013 and has remained active from that time to this. It is likely that SVNL's advantage lies in our more collaborative approach, as the Sim team now has six members who together contribute 10 processors (CPUs and GPUs.) Former Aperture employees describe that private-sector lab as obsessively secretive, with occasional purges of the technical staff in the name of preventing industrial espionage. Stats on team #225029 Aperture having only a single surviving active folder, Dr. Norman Clature.

Yay team!

Congrats on taking over poor Dr. Nomen Clature This all sounds really interesting! I looked over the website and consider doing this, I just have a couple of questions if you don't mind.
- Do you find that it interferes with your general computer activities? And if so, does tuning down the client help?
- Going from the points you guys have collected, do you normally manage to get the process done before the deadline?

I'll just answer from my experience because I don't know about it in a more general way. I have it set to medium and it doesn't interfere with my average activities. It's easy to put it on another setting - I turn it down to idle when I use photoshop or sims 3. I came home the other day and my daughter was playing Mass Effect 2 on this computer with FAH set to Medium and the game was running fine for whatever that's worth.
I'm not sure about deadlines - I think maybe it only matters as far as getting bonus points (?) which wouldn't be that important probably.

My experience agrees with RoseCity's -- on a desktop machine on Medium I do not notice any effect on my usual activities, even on TS3 which has the reputation of being a resource hog. I do hear the fans, though, as the machine is running warmer (I notice this more on Medium or Full-blast, but somewhat even on Light.) I do not know how it impacts a laptop.

As to deadlines: The {tl;dr} answer would be "Yes", at least if a job starts successfully, it is just about certain to be returned within the deadline.

The rest of this post is a more elaborate response to your question about deadlines.
For a concrete example, here is what the FAHControl application says about the job that's on my CPU right now:


If I understand correctly, "Timeout" is the time by which, if my client has not returned the job to the server, the server will withdraw it and assign it to a different machine. That's ten days in the future, whereas under the current conditions I'll probably finish the job in about 19 hours. I suspect (although I don't know for a fact) that the client and server negotiate so that the server knows the client's settings at download time, and preferentially serves Work Units that are likely to complete by the time they're needed. (I do not know whether my habit of playing with the knobs interferes with that negotiation. Now that there's less novelty, I am settling down, and usually leave it on Light or Medium for many days at a time.)

Now as to what RoseCity said about bonus points for a quick turnaround -- here again from my screen:


In this example the Total Credit (which includes the bonus points) is quite a bit higher than the Base Credit, so these quick-turnaround bonuses can evidently be pretty significant. (This justifies taking the slightly perplexing step of obtaining an emailed "Passkey", because the bonus is only awarded to clients that have a passkey at the time of the upload.)

From reading other boards, I gather that the rapid-turnaround bonus is not merely for finishing within the deadline, but for finishing much sooner than the deadline. But even at that, it usually succeeds. I think the effect is sort of visible in our daily stats as we've been ramping up. The points-per-day for each new contributor start out modest, but they appear to jump to a higher level after a few Work Units have been returned over a few days. I think what's happening is that people get their Passkeys configured and their Folding-power knob set comfortably, and then the server starts giving them larger jobs with a high likelihood of being returned fast enough to earn the bonus.

One more thought & caveat: above, I said "if a job starts successfully." Sometimes, for example right now, I find that a job assigned to my GPU fails immediately after downloading, and I don't know why. I have an nVidia card but I have heard informally that this is more likely to happen with Radeon cards. I don't know whether it is specific to the contents of the Work Unit, but maybe it is, because it's happened to me before and then it resolved itself somehow. So that's a factor that can affect your points-per-day but we don't understand quite why. When my GPU is contributing (i.e. most of the time), I can return perhaps three times as much work as I'd get from the CPU alone.

Thank you both for your elaborate replies! Very much appreciated I would be running this on a laptop, the only desktop we have is on XP and I will take that offline when April comes. The laptop is already struggling with TS3, so I would have to keep a close eye on temperature then....

I believe we have one member (Thornton'sLaptop) who is running on one or possibly two laptops. (Sometimes it is hard to tell what's happening through daily reports, because WUs can take more than a day to finish, so the progress is necessarily very lumpy.)

Nysha looks to have run into the same problem as Tsyokawe -- the Stanford stats show multiples of 5 WU's for both members, but zero points. This suggests that a failure was detected only at upload time, and perhaps there is some automatic retry, so that the server counted five uploads but without finally succeeding. Alas! the reason for this is somewhere above my technical skill level.

In all, we now have eight members running twelve processors, but two of the eight are suffering this upload-frustration.

Work Summary
The team has uploaded 338 work units for a total score of 854,054 points. We are steadily turning in about 200k points per week, so we can expect to crack one million points within a week from now.


Presently about 1/14,000 of all the protein folding being done for Stanford is performed by the Twallan SVNL team.

Compared to What?
Of all the teams registered since the beginning of the Folding project, we are now in the 97th percentile by total contribution. Each day we are overtaking about a dozen active teams and five to ten inactive ones (teams that accumulated scores in the past but are no longer turning in WUs.) Rarely are we overtaken by any other teams, although today or tomorrow we will be passed by Team Alex Marien (a team of similar size to ours that recently added another muscular processor.)

Getting into the range that makes a practical difference to the Pande Lab's results, evidently, is mainly a matter of (1) starting to fold and (2) keeping at it.

Upload bug
Two team members (Tsyokawe and Nysha) have got sidelined by a bug that causes their uploads to fail. Nysha goaded her client to re-try but, although Stanford now shows 15 WU (presumably from three attempts at different times), none of them were scored.

@Nysha: do you know whether these upload attempts were all for the same WU, or did the server send you a new one? I am wondering whether the first upload failure caused you to get albatrossed, whereby you can't start a new WU because the old WU never got cleaned up. If that is what happened, you ought to be able to manually remove the work directory of that WU, and see if the server will give you a new one. If it does, and if the new WU is able to finish, then we will know that the trouble is triggered by specific bad WUs, and we will know a workaround.

Today the team earned 52,343 points, nearly twice our daily average. This is mainly the result of @RoseCity turning in the single largest WU ever for our team (33,475), as part of the largest-so-far daily contribution by any member (37,043 points for her four WUs). We are now in the 94th percentile of the approx. 87k active teams (or the 98th percentile of ~219k teams that ever registered.)

I believe I have accidentally duplicated the failure-in-batches-of-5 syndrome that has thwarted @Tsyokawe and @Nysha. Today, I turned in one WU that was scored, and 5 that were not. The failures were on my GPU, and the success was on the CPU (i.e. they were different WU running on different FAHcore images). But the GPU has often folded well for me, and I haven't changed configuration as far as I know. So I would say that we now have some further details and variations on the problem report, but not yet any insight.

Well, it was my CPU and GPU that did all the work - I did squat.

After some supposedly innocent updates, my GPU started doing the same thing as the machines of @Tsyokawe and @Nysha -- it would appear to start a WU, but would fail quickly, with Stanford's server showing 5 WU uploaded and 0 points for them. Cleaning out the relevant work files & re-starting would just do the same thing again. I've found a fix as described below, and am now happily cranking out WUs at a RoseCity-like pace.

Story Time
I have an Nvidia card now, but I used to have an ATI Radeon, and the bug in my case turned out to be what people are calling the "ATI OpenCL driver bug." OpenCL is what FahCore 0x17 uses to achieve a vendor-independent interface to the GPU. OpenCL, accordingly, involves some DLLs that are supposed to be shared amongst different video drivers. The problem is that across vendors and throughout history, drivers do not all agree on where those files are expected to be found or when they're allowed to be removed. You can get stuck with old ones that will thwart your current drivers (from either vendor.) The symptom will be that your graphics card is unable to do anything requiring OpenCL. Well, OpenCL is only one of many possible interfaces and evidently not a hugely popular one, so you might never have seen a problem unless you ran a video card diagnostic tool like GPU Caps , or unless you tried to fold proteins.

How I Got Around It
Truthfully I spazzed around for a while, and it is hard now to say which of the unsuccessful maneuvers might have contributed something to the eventual success. This blog post (due to Lithium) is the closest to putting a successful procedure all in one place:
http://mrlithium.blogspot.com/2013/...driver-bug.html
My own variations/additions to Mr. Lithium's procedure were as follows:

• I now have an Nvidia card, so all the stuff about re-installing AMD Catalyst was inapplicable to me. I knew I would want the current NV drivers for my card, and so I downloaded those into an easy-to-find folder before reaching any steps that made me nervous.
• I downloaded GPU Caps to better see what the card was saying & doing, including specifically its OpenCL situation.
  http://www.geeks3d.com/20140312/gpu...-20-1-released/
• At some point I ran the ATIman Uninstaller, a driver-cleaner specific to ATI Radeon drivers. But I do not know whether that step was either necessary or sufficient! Regardless of whether you use any driver-cleaner, check that the specific OpenCL files named in Lithium's blog post do finally get removed. That is what gives any subsequent driver installation/repair a chance to put the right ones in the right places.

Caveat: This Only Applies to GPUs
All the above is very specific to WUs that FahCore 0x17 loads onto a graphics card. The bug and solution do not apply at all to WUs that run on the host CPU. In fact, even on a machine whose GPU is afflicted by this bug, the host CPU should be able to fold. If anybody is unable to run those jobs (host CPU, usually under FahCore 0xa4), I'd be very interested to hear about it.

As far as I'm aware, my CPU was folding fine - I stopped because I was travelling and couldn't leave my laptop running for the 10+ hours a day it takes to complete a job, not because the process was failing. The info is very interesting, but given the recent problems with TS2 and graphics drivers, I think I'd rather avoid messing with them in order to avoid shadow issues. Hopefully your post will help any TS3 players who don't have to worry about that, though.