Folding@Home; Longevity Team

Good idea, but I think there is so much wasted processing time anyway, that only has to be activated - wasted by only running the "System Idle Process", waiting for the next key stroke.
The effect of such activation could be greater than the effect of 10 new expensive computers. I think many people would gladly join because it costs nothing und not even slows down their own computer, but they not even know that they could help that easy way. (theoretically 4037 forum members, 224 full members, 40 active Folding@Home members in team 32461)

Welcome DRauch.
He had his first Folding@Home-upload today. Thx for joining.

another new warrior: Welcome James (a 100+ debut wow!)

Hi bsmall, Welcome aboard!

Well done, the level-250-quest is finished. So let's fulfil the old predictions and let's reach the 200 !

intermediate result:
http://folding.extre...?s=&t=32461&p=3

yes! Strong reenforcements have arrived.
http://folding.extre....php?s=&t=32461

Sweet!
The statistics are so interesting.

To 200 and beyond!

So how do I add myself, I see a team number entry, but cannot seem to find a number?

chmstar -So how do I add myself, I see a team number entry, but cannot seem to find a number?

Hi chmstar,

The Longevity Meme team number is 32461.

Some F@H info and links to statistics here at The Longevity Meme

Here's a thank you to most recent Longevity Meme F@H Team contributors:

- CG Thomas

- Chad Morningstar (onboard today)


http://folding.extre...?s=&t=32461&p=2

Welcome Paul,

You were the 153rd contributor to sign on with The Longevity Meme F@H team.

http://folding.extre...?s=&t=32461&p=2

Sweet. Up to a rank of 225.

http://folding.extre....php?s=&t=32461

I have been running Rosetta for a few months now. Why is FAH > Rosetta? I left FAH because their application did not easily support dual core CPUs (had to do some funky stuff and run 2 instances at the same time in order to use both CPUs).

You know, Ghostrider, I am not sure, although I am sure Matthias or someone could answer. I think it is probably just because the Folding@Home project is larger and they have team tracking and stuff that there is a longevity team with it. I am not all that familiar with Rosetta, personally, but I am sure they are both worthwhile projects.

As far as the processors, I know that the new Folding@Home software in beta testing right now combines both the CPU and the GPU (graphical processing unit) for what they had claimed would be a 40x improvement in processing, but in the beta test so far has surprised them at about a 70x improvement. (which is really, really impressive) Once they begin putting this into all the software, it should really jump up the amount of work units being produced.

It really is amazing how there are side technologies like this (the massive improvements in ways to process data efficiently on an individual computer) that can come as a byproduct of trying to figure out how proteins fold. Perhaps I am just easily impressed..

Also, I think they have come out with a client that runs on the Playstation 3, which is kind of cool.

http://en.wikipedia....ki/Folding@home

Ghostrider, here is a quote from the Rosetta@Home wikipedia page on the difference between the two projects.

Says Dr. Vijay S. Pande (of the F@H team),

    I know Baker and Ranganathan and their work very well and (like the rest of the protein community) find their work very important and impressive. However, Rosetta@home and Folding@Home are addressing very different problems.

    Rosetta only predicts the final folded state, not how do proteins fold (and Rosetta has nothing to do with protein misfolding). Thus, those methods are not useful for the questions we're interested in and the diseases we're tackling (Alzheimer's Disease and other aggregation related diseases).

    Also, one should note that accurate computational protein structure prediction is still very challenging compared to what one can do experimentally, whereas the information obtained from Folding@home on the nature of folding and misfolding pathways matches experiment (e.g. with quantitative validation in rates, free energy, etc) and then goes beyond what experiment can tell us in that arena. While Rosetta has gone a long way and is a very impressive project, given the choice between a Rosetta predicted structure and a crystal structure, one would always chose the crystal structure. I bet that will be changing due to their great efforts, but that may still be a ways off for that dream to be realized.

    So, both are valuable projects IMHO, but addressing very different questions. I think there are some misunderstandings out there, though. Some people think FAH is all about structure prediction (which it is not -- that's Rosetta's strength) and some think Rosetta is about misfolding related disease (which it's not, that's Folding@Home's strength). Hopefully this post helps straighten some of that out.

Ghostrider -...FAH because their application did not easily support dual core CPUs (had to do some funky stuff and run 2 instances at the same time in order to use both CPUs).

Hi Ghostrider

You are correct, curently one must run either the Linux or Mac OS (Intel) F@H cliet for SMP support (multi-cores). Although, Stanford is set to release an SMP client for Windows pretty soon.

Runnning two instance with dual-core works fine. WinXP-Pro can handle many instances. This allows a dual-core computer to perform almost equivalent to two computers, with two Work Units processed simultaneously.

Of course there is not much to be gained by running more than one instance per CPU core, it almost always lowers efficiency. Unless, one is running a Pentium 4 with hyperthreading, then running two instances per core can improve the processor efficiency.

Stanford F@H is offering good bonus points to the contributors running the SMP client (currently Linux or OSX). If you are running Windows, watch the download page for the upcoming SMP client:

http://folding.stanf...u/download.html

Currently, the Windows SMP client is in open beta test. If you want, you can download the beta version now.:

I liked the FAH project overall. Besides the lack of SMP support, part of the reason that I left is because they had such a large audience (of course it is not large enough), and they were doing cool things to squeeze computing power out of GPUs. I left to Rosetta because the opinion at betterhumans.com at the time was that Rosetta looked like the best shot for longevity research and FAH had more resources and computing power to throw at the problem. I felt like my cycles would make more of a difference at Rosetta which does support SMP through their Bionic client. Of course, if I can be convinced that FAH really is the better approach and that my computing cycles can be put to better use there, I will switch back.

Ghostrider-

I liked the FAH project overall….a large audience (of course it is not large enough),

Agreed, Ghostrider.

Stanford has a good number of contributors (about 200,000 active), but this is not nearly enough. Considering that there are about a billion internet users worldwide, it is a small proportion (about 1 contributor per 5,000).

There is a lot of untapped potential here. If every person that has the capability were to set their computer to contributing, then progress in this area of research could speed up by more than 1,000 times the current rate.

Where else can the rate of progress be influenced by such orders of magnitude?

….and they were doing cool things to squeeze computing power out of GPUs.

Yes, GPUs and PS3s.

I suspect that the largest groups of contributors are the gamers and computer hobbyists, like [H]ardOCP and Overclockers Australia, etc.

…if I can be convinced that FAH really is the better approach and that my computing cycles can be put to better use there, I will switch back.

Contribute where you think best.

Ghostrider, that you contribute to advancing the cause is commendable.

Welcome bwille24,

You were the 154th contributor to sign on with The Longevity Meme F@H team.

Already processed two Work Units, listed here:

http://vspx27.stanfo...e&teamnum=32461

Currently, not yet listed at EOC, here:

http://folding.extre...?s=&t=32461&p=2

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

And, good news reported today on PS3 contributions to F@H in this thread:

Edited by dnamechanic, 24 March 2007 - 02:51 PM.

Funny thing, I don't know what exactly took me so long to notice this, but perhaps it was a recent Digg entry that mentioned how much Folding@Home was being accomplished on Playstations of all things.

Anyway, I've got a long way to go to catch up, but hopefully I can figure out how to utilize my dual-core effectively. I've always struggled with distributed computing (having done SETI@Home in the past) because it always seems like it's not using enough of my available resources. A program that automatically detects dual processors and maximizes the usage of the CPUs and memory when my machine is idle would be very handy.

This is somewhat unrelated, but kind of relevant to the current thread :-). A while ago I upgraded one of my computers, a Celeron 2.0 GHz (478-pin) system after the motherboard died (power supply killed it). I still have the MB (which is dead, but may be repairable -- somehow), CPU and fan / heatsink. The CPU should (although I am not completely sure) still be good if anyone wants it. Pay for shipping and I can send it to you. The CPU is pretty much bottom end, but could provide a cheap upgrade or second computer for someone who happens to have a bare P4 board lying around. I can send the board too, but I think it is dead. PM if interested. I would like the CPU to be used for folding.

-JMorgan... program that automatically detects dual processors and maximizes the usage of the CPUs and memory when my machine is idle would be very handy.

Running two instances of the Folding client works fine. Each FAH client will fully utlize a core of the CPU. This seems to work very well; most contributors with dual core systems use this approach.

Also, the FAH client defaults to lowest priority so that any other program running (at normal priority) will not be slowed. If desired, there options in the configuration that allows flexibility in usage, although for most contributors the default settings are fine.

As mentioned earlier, the FAH Windows SMP client will be available soon. The SMP client allows multiple cores to work on the same FAH Work Unit. This is advantageous for Stanford, in that the turnaround time for a given Work Unit is substantially decreased. Probably for this reason, additional bonus points are awarded for the completed SMP Work Units.

Welcome two new members to The Longevity Meme team:

- C_Philip

- tweaver113

You were the 155th and 156th contributors to sign on.

Thanks

I haven't been able to use my folding software for months! My version won't connect to the server for some reason... I've completely deleted it and re-installed it, all that jazz, but the mac graphical version may be getting too little love.

Ghostrider -    I upgraded one of my computers, a Celeron 2.0 GHz (478-pin) system after the motherboard died (power supply killed it).  I still have the MB (which is dead, but may be repairable -- somehow), CPU and fan / heatsink.  The CPU should (although I am not completely sure) still be good if anyone wants it.  Pay for shipping and I can send it to you. 

Ghostrider, if you do not get a response here at ImmInst.

You could post a listing in the For Sale or Trade sections in any of these forums:

http://www.hardforum...isplay.php?f=17

or here,

http://forums.anandt...s.aspx?catid=45

or here:

http://forums.pcper....display.php?f=3

Or, maybe try posting a statement that the hardware is available, with a link to this thread, in one or all those forums. The link might be helpful in other ways.

I would like the CPU to be used for folding.

Often, someone is looking for hardware items to use in folding rigs at these forums.

Wow. I hope all the work units aren't this long. Even with a very decent machine, this one seems to be taking forever. The client says it's on 10600 out of 20000 frames and should be done in another 2 days.

I've looked on the folding@home forum about the discrepancy between how the client shows 20000 frames, yet their site shows this particular protein as having 100 frames. Is there a reason for this? How long do your work units take on average?

Ghostrider, if you do not get a response here at ImmInst.

You could post a listing in the For Sale or Trade sections in any of these forums:

http://www.hardforum...isplay.php?f=17

or here,

http://forums.anandt...s.aspx?catid=45

or here:

http://forums.pcper....display.php?f=3

Or, maybe try posting a statement that the hardware is available, with a link to this thread, in one or all those forums.  The link might be helpful in other ways.


Often, someone is looking for hardware items to use in folding rigs at these forums.

Yup, I know. Posted there too, one guy contacted me and might be interesting depending on if he can get a MB.

JMorgan - Wow.  I hope all the work units aren't this long.  Even with a very decent machine, this one seems to be taking forever.  The client says it's on 10600 out of 20000 frames and should be done in another 2 days.

Likely, you are folding a Gromacs work unit.

  I've looked on the folding@home forum about the discrepancy between how the client shows 20000 frames, yet their site shows this particular protein as having 100 frames. Is there a reason for this?

As far as I know all proteins currently issued have 100 frames per Work Unit. Several proteins are broken into steps, some have 20,000,000 steps with 200,000 steps per frame, but they still 100 frames. When Stanford was issuing Proteins that fold on a Tinker core, many of those had 400 frames. (Edit - I checked Stanford's Projects, they are still issuing Tinker work units and they are 400 frames.)

How long do your work units take on average?

The Stanford performance benchmark for the p21xx Work Units (Gromacs) are about 3 to 3.5 days to fold on a on a 2.8 GHz Pentium 4 machine.

The point value of the particular protein provides a good indicator of how long it will take to fold. The standard benchmark (as I recall) is 110 points per day on a 2.8 GHz P-4 machine. For example, a standard Gromacs protein that has 396 points would require about 3.6 days to fold. That is, if one were using a 2.8 GHz P-4. On other machines it could take more or less time.

The newer Core Duo and Core 2 Duo CPUs fold considerably more per GHz than the benchmark P-4s. On older desktop CPUs (a year or so old), AMD CPUs typically fold faster, on a per GHz basis, than Intel P-4s. As for IBM CPUs in Mac computers, G5s do pretty well per GHz, and G4s are very slow. The Intel CPUs in newer Mac machines are based on the Core architectures, so they are very good folders.

Of course there are other issues, multiple-core CPUs and SMP, GPUs, PS3s etc. All these use different F@H processing cores.

The word "core" as used in FAH context has two meanings. One "core" is the number of hardware processing units in a central processing unit (CPU) such as dual-core. The other of the word "core", refers to the software program that performs the calculations; Tinker, Gromacs, Double-Gromacs, GBGromacs, Amber, Gromacs SREM, QMD, etc.

Edited by dnamechanic, 28 March 2007 - 02:26 AM.

Joel Morgan, welcome to The Longevity Meme team:

You were the 157th contributor to sign on.

http://folding.extre...?s=&t=32461&p=2

Thanks

2ch@PS3 wow!, this shooting star will at least become the new no.5 of F@H
http://folding.extre...?s=&t=32461&p=3