Posts by Paydirt

1) Message boards : Number crunching : teraFLOPS estimate? (Message 53958)
Posted 24 Jun 2008 by Profile Paydirt
Post:
while I hope I'm wrong, it's possible that folks are leaving for Folding@Home. F@H just added nVidia cards and there is considerable excitement around F@H. They are now getting over 700 teraflops from GPUs alone (not counting PS3)
2) Message boards : Number crunching : F@H adds nVidia GPUs to crunch on (Message 53845)
Posted 19 Jun 2008 by Profile Paydirt
Post:
F@H's open beta for nVidia GPUs has begun. In the past 3 days, they've added roughly 2000 GPUs and 300 teraflops. If that is all from nVidia GPUs (as opposed to some of the teraflop increase coming from improvements to the ATI code), that means the average nVidia GPU is doing 150 gigaflops!

The new 4870 (and maybe 4850?) from ATI should be pretty powerful too (ATI is claiming 200 gigaflops of double precision, 1 teraflop of single precision)

Currently: 442 teraflops, 4021 GPUs (both ATI & nVidia), and rising.

http://fah-web.stanford.edu/cgi-bin/main.py?qtype=osstats


With over 2 petaflops of processing power, I wonder how long it will take for them to have a breakthru with Alzheimer's?
3) Message boards : Number crunching : Xeon processors produce less credits? (Message 51953)
Posted 15 Mar 2008 by Profile Paydirt
Post:
Yeah you can do that too with VM.
4) Message boards : Number crunching : Not getting enough work... (Message 51887)
Posted 10 Mar 2008 by Profile Paydirt
Post:
If you can also set the setting for "Connect Every" to 1 day or less (if it is almost always hooked onto a network), that would also induce it to download more work.
5) Message boards : Number crunching : Xeon processors produce less credits? (Message 51886)
Posted 10 Mar 2008 by Profile Paydirt
Post:
I have to agree about HT. Distributed computing uses 100% of each core. So when you do HT to have each core do two WUs, then they are going to be competing for CPU memory cache. There's no magic thing where a program that uses 100% of the CPU will let you run two programs and use 200% of the CPU.

A simplistic way to view HT and VM is that they add efficiency when non-intensive programs run on them. The big idea behind VM is if you have an office with 4 people on the internet and in word processors, you could run them all on one CPU instead of 4 CPUs.
6) Message boards : Number crunching : Mac Pro, 8 cores (Message 51847)
Posted 7 Mar 2008 by Profile Paydirt
Post:
And its not just about an increase in processing power, but also save a decent amount of electrical power and space to get rid of those old machines.

An extreme example would be F@H's GPU folding... The newer ATI GPUs may do TWICE the work of the old ones and use 50% less power (while folding). Over the course of the year, that $180 card could save $100/year in electricity costs. So the card pays for itself!
7) Message boards : Number crunching : Xeon processors produce less credits? (Message 51765)
Posted 2 Mar 2008 by Profile Paydirt
Post:
R@H is affected by memory performance. So I would try doing 2 workunits without the virtual machine? RAM is also important, do both Xeon's have 2-slot DDR2? or 4-slot FB-DIMMs (even better)?
8) Message boards : Number crunching : Computer array crunching? (Message 51725)
Posted 29 Feb 2008 by Profile Paydirt
Post:
PS3 vs GPU

PS3s would be an easy way to go, though I am unsure how they respond (if anything changes) when no display is attached. 0% of getting a discount.

GPU folding is much more complex and is twice as effective as PS3 when you count both equipment and power costs. The 2x might become 8x whenever the FAH GPU client for the ATI 3870 (X2) comes out. This is because the 3870 may only eat up 100W while folding (saving $100/year in electric bills), whereas the 1950XT eats up 200W. GPU folding is much more demanding on the user to maintain, at present. I do not recommend dual-slot GPU folding, because the systems are much more prone to fail as far as scientific folding is concerned.


R@H vs F@H

R@H is more concerned about predicting the final structure and interactions between proteins (or proteins and drug candidates). I respect the methodology of R@H more. F@H is more concerned with the process of protein folding and mis-folding. It is believed that Alzheimer's is caused by mis-folding proteins.

(feel free to correct my synopsi)
9) Message boards : Number crunching : Sponsored Computer for anyone interested (Message 51724)
Posted 29 Feb 2008 by Profile Paydirt
Post:
he sent me a refund saying that his ISP is cracking down on his hub...
10) Message boards : Number crunching : Sponsored Computer for anyone interested (Message 51701)
Posted 27 Feb 2008 by Profile Paydirt
Post:
buyer beware, I donated to chungenhung over 48 hours ago. I haven't heard from him yet. I will keep everyone posted.
11) Message boards : Number crunching : Computer array crunching? (Message 51684)
Posted 27 Feb 2008 by Profile Paydirt
Post:
I would search either the Folding@Home forums or SETI forums for "headless" or "diskless". F@H forums may have been wiped since the big post about this, but if you inquired there, I'm sure people could point you in the right direction.

You could also search for diskless on google. This yielded this site (which was last updated in 2002, hehe, but is for diskless Linux):
http://www.schnozzle.org/~coldwell/diskless/

F@H also has monitoring software so you could know if a node goes down (if crunching for F@H). It was called FAHMON.

Anyhow, keep us posted on what you do and how it goes!
12) Message boards : Rosetta@home Science : How can we bring more users to the Rosetta project? (Message 51629)
Posted 25 Feb 2008 by Profile Paydirt
Post:
Eric, I like your ideas except for purchasing advertising. I think the following ideas would be the most successful:

1) Members recruiting friends, family, co-workers
2) E-mail newsletter would be cost-effective ("thanks for contributing to the project. Here's what we are up to now." even if there isn't much news they could just remind folks they could use all the computing power they can get)
3) More YouTube videos talking about different aspects of the research, but still in layman's terms. (maybe we could brainstorm topics? "chemistry [predictive] vs biology [descriptive]")
4) Educational programs (getting a school to crunch would rock!)
5) Press availability, story ideas for the press (story ideas meaning, if the press can't think of a story, maybe we can think of a story for them, doing some of the work for them).

13) Message boards : Number crunching : BOINC distributed computing criticisms / rebuttals (Message 51626)
Posted 25 Feb 2008 by Profile Paydirt
Post:
I think current contributors can tend to get anxious especially if they haven't heard that a scientific breakthrough has occurred yet. Most of the communication from the project is old, and we simply must trust that progress is continuing to be made. I don't see why suddenly work would stop on the project? It's basically the life's work of at least several scientists at the University of Washington.

As far as power usage, that argument just doesn't fly with me AT ALL. For all the things that you can use power for, I would think "searching for a cure to diseases," or "taking a quantum leap forward in protein research." is a pretty good use of money/carbon. I dunno, I could be using my computer to play games, surf the net, or I could turn it off.

The only thing your team member lists that DOES hurt to idle is the graphics card, everything else doesn't spend much power while idling...

If suddenly millions of computers crunched for science instead of idling, I don't think it would make a dent in pollution. AND if that was a case, we'd be much closer to a cure and we wouldn't NEED to crunch for science for an infinite amount of time. We COULD stop or pull back.
14) Message boards : Number crunching : Seriously, what for? (Message 51624)
Posted 25 Feb 2008 by Profile Paydirt
Post:
What Rosetta mainly studies (as far as I understand) is the 3D shape that proteins take when they fold. Besides using expensive crystalline x-ray photography (and possibly computer simulations), it is not known what exact shapes that proteins take in nature. Nor is it known how proteins interact exactly. With enough computational power we can gain this knowledge and produce drugs to treat diseases with much greater ease AND speed.

Otherwise, scientists are swimming in the dark.
15) Message boards : Number crunching : Seriously, what for? (Message 51621)
Posted 25 Feb 2008 by Profile Paydirt
Post:
Oh yeah, the YouTube video is a cool start, but maybe it was just more powerful for me because I already understood the project.

http://www.youtube.com/watch?v=GzATbET3g54
16) Message boards : Number crunching : Seriously, what for? (Message 51620)
Posted 25 Feb 2008 by Profile Paydirt
Post:
you can't tax deduct 'crunching for science'.

Anyways, Paul this is kinda what I said to another person who knew about proteins... Chemistry is a predictive science, you have chemicals that go into a reaction and you can precisely measure and know exactly what will result from the reaction. Biology is currently a descriptive science because of all the complexities. It is widely believed that with enough computer power and research into proteins, etc, that biology could become a predictive science. There is not going to be a major breakthrough until enough computer power is tackling the problem or the computers have had enough time working on a problem. It is also widely believed that many diseases are the result of proteins "misbehaving" and if proteins can be studied enough then we could have a breakthrough with a disease.

A year ago, Rosetta@Home had about 35 teraflops of computing power, it now has over 63 teraflops of computing power.

As far as what the project has done to date (besides what Mod.Sense states), it almost sounds like a trust issue. Some people look at trust as far as, "You have to show me that you are trustworthy before I can trust you." This is bringing an energy of "mistrust" to the situation by forcing people to PROVE they are trustworthy. If she were a trusting person, she might be more receptive (instead of challenging you) and trust your judgment. Or, maybe she just wants to know more and you could tell her about the diseases R@H is working on or swing her over to select parts of the website/articles.

Sorry about the tangent, people who withhold their trust still think they are trusting people. There is a difference between making people prove trustworthiness and simply being a trusting person (but not being stupid about it). The thing is that people sense when you are a "prove it" person and it affects the energy you bring to people when you interact with them.


I think people also get impatient about giving some times. They want results now. What if your wife gave to women's shelters but their clients continued to grow or continued to be abused. Or what if she simply proposed it and you replied, "What makes them special compared to all the other women's shelters out there?" Seems like a crappy attitude to take.
17) Message boards : Number crunching : Sponsored Computer for anyone interested (Message 51617)
Posted 25 Feb 2008 by Profile Paydirt
Post:
COST?

To determine the full cost, you gotta figure in a life-span for the equipment (I conservatively use 3 years) and divide the equipment costs by total months (36) and then add the monthly electricity bill to figure out a monthly cost that includes equipment and electricity (but donates your time). Space also has a cost depending on how big you go.


NON-PROFIT?

I looked into whether or not I could create a formal charity that crunched for science. The U.S. government does not currently recognize 'computer farms crunching numbers for science research' as something that can gain non-profit status. I did not apply myself to become a charity, but numerous SETI folks have tried to do this.


TAXES?

If you want to receive monies from other people for a hobby (since it can't be a charity), then you will need to show the cost of your hobby is greater than the money you take in, but you can only do this for the period of 3 years (have a non-profitable hobby). After then, if you want to be in compliance with Federal Law, you would have to do more relating to taxes (I am not an accountant). So, if you did this for more than 3 years, you may be setting yourself up to be audited by the IRS.
18) Message boards : Number crunching : Seriously, what for? (Message 51505)
Posted 19 Feb 2008 by Profile Paydirt
Post:
Nidhogg,

You are making a donation. Donating is supposed to be a selfless act, yet you want everyone to pat you on the back?

I think you are getting a little overly anxious. You know that you are crunching 24/7 for a project, what more do you need?

If you are looking for stats, because you are fascinated by how the numbers change, search for "boinc bam", then you can find your user name or e-mail or whatnot. I don't think it tracks WUs though.



Paul,

Yeah PS3 is currently the best bang for the buck with the least setup time. However, per dollar spent (equipment and power costs), crunching F@H with a ATI x1950 is TWICE as powerful. All you need is 16x PCIe and a CPU core that is not utilized for anything (the GPU ties up a CPU core), the CPU does not need to be powerful. GPU folding is more finicky and requires a little more monitoring though.

Anyways, I appreciate what you are doing and the commitment you've made. I intend to spend 5% of my paycheck on crunching/folding machines. I'm waiting for a F@H client that can utilize the ATI R2XXX and R3XXX GPUs. I am running 2 GPUs (and have 1 GPU collecting dust) for F@H and various CPUs running R@H.
19) Message boards : Number crunching : OMFG!!!!!!! folding just broke the teraflop boundrie!!!!! (Message 47356)
Posted 3 Oct 2007 by Profile Paydirt
Post:
The comparison doesn't really need to be redone.

Here's why: With the $100 drop in the PS3, the ATI GPU also had a price drop of $100. Before the latest PS3 drop, the GPU cruncher has 2x the performance per dollar spent (over 3 years, counting power costs). While another PS3 price drop is a good thing, the GPU cruncher would still beat it by a factor of around 80% (more performance per dollar spent)
20) Message boards : Number crunching : OMFG!!!!!!! folding just broke the teraflop boundrie!!!!! (Message 47250)
Posted 30 Sep 2007 by Profile Paydirt
Post:
Running Folding@Home on PS3 is mostly "fire and forget" like BOINC. Doing PS3GRID on PS3 is much more hassle (you have to install linux, etc).

GPU folding is still in its beta stage and it is difficult to set up on a computer. You have to use specific drivers (you can't always use the latest drivers). There's a limit to how much you overclock, but with x1950xt cards you have to clock your card because the Folding application does not trigger the card to run in standard 3d-gaming clock speeds. So you have to use a secondary program to increase core GPU clock, give it enough voltage at those clocks and run the fan faster than standard (60% flat fan speed is good). It's also good to have a GPU card that is a "2-slot" card that spits the heat directly out of the PC and have an intake vent on your computer case that is directly over the GPU. (Not so simple as buying a PS3, downloading the app, and then running the app) :)

I do think GPU folding is worth it because it has twice the performance of PS3 folding PER dollar spent over 3 years (including 3 years of power costs).


Next 20



©2021 University of Washington
https://www.bakerlab.org