Talk:Folding@home/Archive 2

Archive 1

Archive 2

Archive 3

Results (and better summarisation of them)

I believe it's important that results be better summarized in lay language or in such a manner that they explain better what the results are all about and what they may lead to. -Mardus (talk) 15:15, 28 April 2008 (UTC)

Agreeing with the above. I came here to see what had been achieved by the project, but this long string of scientific manuscripts with names of authors etc isn't really helping. Presumably someone associated with the project could provide a brief synopsis of the most significant developments in research that this project has helped to bring about. Ubertoaster (talk) 08:24, 19 April 2011 (UTC)

See this http://folding.stanford.edu/English/FAQ and particularly this http://folding.stanford.edu/English/FAQ-Diseases. I'm participating in the project, and am pretty passionate about it. There are many others out there, but I can certainly help you. On the suggestion, I'm considering moving the list of publications to a separate article. Also, the cores need to be dealt with, perhaps in a similar manner. Thoughts? Jessemv (talk) 00:17, 21 August 2011 (UTC)

Second biggest

Folding@home has now been surpassed in terms of cumputer power by bitcoin. According to bitcoinwatch.com, bitcoin has a total power of about 19 Petaflop/s. The article's claim of being the biggest distributed computing cluster is thus false. 213.66.122.5 (talk) 10:07, 12 May 2011 (UTC)

Is it even a distributed computing project? According to the website, it's simply a currency transaction/trading application.

194.80.64.113 (talk) 15:53, 12 July 2011 (UTC)

It's a glorified pyramid scheme from where I see it 122.57.186.20 (talk) 09:25, 3 August 2011 (UTC)

I don't think BitCoin is a distributed-computing project. Its just a money exchanger thing, and the petaFLOPS don't really mean anything other than that stuff is moving around. If I recall right, you have to do some brief calculations to perform a trade, or something like that. So its purpose is trading, not computing. The computing is a side effect. Besides, the article has been changed by someone else to no longer claim that it is the largest. IMHO, it is though. Jessemv (talk) 00:25, 21 August 2011 (UTC)

Suggestions for GA

I noticed this as a GA candidate, and after glancing at the article, if I were the reviewer, I'd fail it due to the extremely long embedded lists (part of criteria 1a of good article criteria). Quite frankly, the reader is probably not interested in the details about every single Active/Inactive Core, nor is the reader interested in every single scholarly publication that has used Folding@home data. You might want to split these lists out into their own sub-articles, and write a paragraph or two of prose to summarize the data instead. ...comments? ~BFizz 16:40, 16 August 2010 (UTC)

Fixed a while back. Better now? GA process is likely to start up again. The last attempts at GA and FA were quite frankly pitiful and embarrassing. Unlike those attempts, I am honestly looking to improve the article to GA/FA status, and it is ALMOST there. Standby. Jessemv (talk) 20:06, 24 September 2011 (UTC)

GA Review

This review is transcluded from Talk:Folding@home/GA1. The edit link for this section can be used to add comments to the review.

Reviewer: Wizardman _{Operation Big Bear} 15:02, 17 August 2010 (UTC)

After a cursory read, I see that this article meets quick-fail criteria, as a result of much of the article being embedded lists, the references being bare URLs, and other issues. As the nominator is indef-blocked, there's no one to fix it, so I'm failing the article. Wizardman _{Operation Big Bear} 15:02, 17 August 2010 (UTC)

Why isn't there an update on the Petaflops achievment.

I realize that there are two measurements of the FLOPS. The nativeFLOPS and the x86FLOPS But I was wondering why the achievement of reaching sustaining 6.2 x86FLOPS on April 2010 isn't mentioned or some sort of comparable graph to the between the two. As it is credited as being one of the records on FLOPS page but not on here. —Preceding unsigned comment added by 71.59.218.245 (talk) 17:34, 1 March 2011 (UTC)

That issue should be fixed now, but not exactly by a graph. See the native FLOPS are used because those records go back. We simply have no information as to the x86 FLOPS way back in some of the early milestones. The article mentions the conversions, and if the reader wants more info they can click on the citation. Let me know if you have any further suggestions. Jessemv (talk) 20:09, 24 September 2011 (UTC)

Wikipedia team

Just wanted to point out that the wikipedia team # is 42223 - CompuTerror™ 13:39, 5 August 2011 (UTC)

Thanks I guess. There's a link to the team under External Links. Jessemv (talk) 20:11, 24 September 2011 (UTC)

Folding@home cores

I have recently moved the "Cores" section of the article to its own page. Here are some of my reasons for doing so: After reviewing policies regarding what can and cannot be in an article, I felt that the Cores section should be its own article. Here are some of my reasons: 1) At it was in the F@H article, the list was very long, and annoying to scroll through. 2) I don't see the list getting shorter. To the contrary, as new cores are added and old ones retire, the list will grow. Thus it should be split off somehow. 3) For those coming to the F@h article, it is likely irrelevant information. They end up having to scroll through this big monotonous list to get to the rest of the article. Most people probably look up F@h on Wikipedia not for its cores, but for general information about the project itself, what it has done, and where it is going. This is covered very well in the article minus the Cores section. 4) There is a lot more information specifically about cores than what was currently there. What was there is little notes and things. I believe that people didn't expand on it for reasons including that it would worsen the length of the list. Also its a bit technical, but some people have the know-how to add things, but there's a lot of general information that can be added as well. 5) The cores have enough references and enough to talk about to warrant their own article. Certainly information specifically about GROMACS or TINKER could be brought in. Information on specific cores is found in other places. Its not a main talking point here on the forums, but if one hunts around some good information can be found. There's probably a lot more information about them than can be found for some other Wikipedia articles, like the list of all obscure fictional Jedi who ever lived. 6) I or someone else can turn the list into paragraphs that flow well and are easier to read. This would also make it more encyclopedic. 7) Wikipedia's Be Bold policy. 8) There was a topic on this talk page about the length of the Core section and how they didn't think it belonged. It is important information, so I didn't want to get rid of it altogether.

I hope I referenced the article all right, but of course feel free to change it. The Folding@home cores article is fairly new, and needs some cleanup. Also, I don't think I'm allowed to remove the "New Article" tag, so perhaps someone else could eventually replace it with a Stub tag or something. Finally, if you can, please edit and expand it! Thanks. Jessemv (talk) 00:18, 22 August 2011 (UTC)

New Article flag has been removed. Would be nice if the article had some more information to each core though. Jessemv (talk) 15:04, 20 September 2011 (UTC)

Publications

Surely this article should,have a list of publications or summary of results which have come from the Folding@Home project? Publication is an important part of any scientific research and arguably the reason for doing it (in that the whole purpose of research is to present new results). It would be really neat if F@H participants could read the Wikipedia article, see the results and think "Hey, I helped do that". Non participants would no doubt naturally ask the question "why" as they read the article? Results are part of the reason why F@H does what it does. John Dalton 23:36, 4 November 2007 (UTC)

The third paragraph mentions it. Are you saying that it should have a more prominent discussion on the papers? (like a full section) If there have been any major breakthroughs that have resulted, then I would tend to agree with you. Cardsplayer4life 23:53, 4 November 2007 (UTC)

I agree. I'm quite curious what specific scientific knowledge has been gleaned from these simulations. Also, I think a mention of how the volunteers are credited for their work in these scientific papers should also be there. --seav (talk) 11:08, 21 November 2007 (UTC)

Although it needs a fair amount of work, I have added a list of published results to the article. Johnnaylor (talk) 19:36, 29 February 2008 (UTC)

I'm starting work on this summary. You are clearly right. Jonnaylor's list of publications was nice for their sheer numbers, but summary style is much better. It's going to be a lot of work, and may take me a while, but I just wanted to let everyone know that I'm currently working on it. My goal is to get this article to a Good Article or near it by January, when the v7 client will become the recommended client and the F@h website gets overhauled. Jessemv (talk) 20:45, 9 October 2011 (UTC)

Estimated energy use and efficiency

I'd just like to point out that I kind of don't like this section, but I'm not positive that we should get rid of it. I really see two options here. One, I could delete the section entirely. Two, I could copy the section onto my own website, reword to make it look better, and then cite that page in this article. Pros for option one: no other distributed-computing page has this analysis, its difficult to estimate anyway, and its probably irrelevent. Cons for option one: some people would like to know. I have seen several people on the F@h forums referencing that section, so I know that it does matter to those people. Pros for option two: keeps the apparently desired content, and removes the article of Original Research. Cons for option two: that's kind of a weird thing to do, plus the estimate is still going to be a shot in the dark, and it may likely end up not sounding like the rest of the article. So those are my thoughts. What do you think? Jessemv (talk) 15:02, 20 September 2011 (UTC)

And by the way, would someone please fix those URLs in the Results section. I'm not sure how to do it properly, but it doesn't look good having some of those recent papers just be external links. Thanks. Jessemv (talk) 15:02, 20 September 2011 (UTC)

Correction: URLs fixed. I tried again and I guess it was easier to get all that information than I thought. I still couldn't get as much citation details as I would have liked, but I tried my best there. Please feel free to add more! Jessemv (talk) 02:01, 22 September 2011 (UTC)

By the way, just discovered that that section dates back to February 24, 2008. Is it really important? I feel like we should make a decision on that, and then I'd be happy to learn more of Wikipedia's workings by nominating this as a Good Article, since it seems to me that other than that section it is worthy, or is a few minor edits away from it. Hopefully we can get it to Featured. If Rosetta@home's article is Featured, ours certainly should be! :) Jessemv (talk) 02:01, 22 September 2011 (UTC)

I am in the process of trying to find external analysis of this. Neither PrimeGrid, Rosetta@home, SETI@home, or World Community Grid have this section in their articles. Nevertheless I'll see if I can get some non-OR content in there. If it becomes reasonable to assume that nothing can be found, I move to delete the section. I am happy to do so if no one has any objections. Thoughts? Jessemv (talk) 17:19, 22 September 2011 (UTC)

I have not yet found legitimate analysis. I have Private Messenged several high-ranking users (one a Site Admin) on foldingforum.org, and one said, among other things, "At this point I'd probably recommend that you leave out that section, for the reasons you suggest." and reminded me that "The resources that FAH uses generally don't include every possible type of computing component". The other user replied back with "It can go away as far as I am concerned. Without citations, it really doesn't belong." So, the section currently has three delete votes (counting mine at this point unless I can remove the OR) and external analysis has yet to be found. In about a week, I will be posting my intentions, waiting a week or two for responses, and then proceeding with deletion or whatever. Just FYI everyone. Jessemv (talk) 06:51, 24 September 2011 (UTC)

Alright. Well apparently Dr. Vijay Pande has not gotten back to me with any energy consumption references. I guess he's too busy studying all our WUs. :) So, it sits with two votes against from high-ranking forum users, and my down vote as well. But I've decided not to completely delete the section. Just so no one freaks out because they cared about it and finds it's suddenly gone, it's right here, listed below. Hopefully now we can go improve the rest of the article, and then I'll see about getting on with the Good Article nomination process. *chorus* Still, if anyone finds any legitimate, comprehensive, and appropriate external analysis feel free to apply it and put the section back in the article. In the meantime, here it is, but without the September 2011 Original Research flag. Jessemv (talk) 17:42, 28 September 2011 (UTC)

Folding@home is a diverse network, utilizing many different kinds of CPUs, GPUs, and PS3s, and in some cases is not the exclusive usage of the system. This wide range of hardware and the nature of distributed computing makes calculating energy use difficult, and requires a number of assumptions. However, some estimates can still be made.

Starting with the assumption that the average desktop computer draws about 65 to 250 watts, and the average laptop draws at most around 50 watts, then we have a CPU energy usage of 50 to 250 watts. As of August 9, 2011 there were 403,858 active CPU clients providing 643 teraFLOPS (643,000,000 megaFLOPS) of computing power. In the worst-case scenario that there is only one CPU client per computer, we have a total between 20,192,900 and 100,964,500 watts of power for a efficiency between of 6 to 32 megaFLOPS/watt.

The GPU Folding@home client can be run alongside the CPU client, or alone. We can assume that the power draw of the computer doubles when the graphics card is fully utilized. Note that this assumption also includes the possibility of the GPU client being run alone. As of August 9, 2011 there were 18,604 ATI and nVidia GPUs contributing 4,446 x86 teraFLOPS (4,446,000,000 x86 megaFLOPS) of computing power. Thus assuming a power draw between 50 and 250 watts each, this amounts to between 930,200 and 4,651,000 watts total, translating to between 956 and 4,780 megaFLOPS/watt. At worst, this would put Folding@home's GPUs 5th in the June 2011 Green500 List.

Official Folding@home estimates that each PS3 will draw 200 watts while running Folding@home. As of August 9, 2011, 19,803 PS3s were providing 1,177 x86 teraFLOPS (1,177,000,000 x86 megaFLOPS) of computing power. However, the newer PS3s draw about 90 watts while folding. Assuming 90 watts each, it comes to a total of 1,782,270 watts for an efficiency of 660 megaFLOPS/watt. This would place Folding@home's PS3s 12th in the June 2011 Green500 List.

With these power-usage assumptions, this amounts to a total between 22,905,370 and 109,576,100 watts, and with 6,266,000,000 megaFLOPS overall, this is an average of between 57 and 273 megaFLOPS/watt, placing Folding@home between 87 and 421st in the June 2011 Green500 List. This only takes into consideration running the computing system itself, and does not factor in heat production, normal system load, nor dynamic frequency scaling.

Jessemv (talk) 17:42, 28 September 2011 (UTC)

Images

I am aware that this article needs some more images. I'm working on it, but the biggest thing is getting permissions and whatnot. I've got some really great illustrations picked out. Recently I have commented out that one image of F@h using 99% of a CPU. IMO, it wasn't very helpful (as its been explained in words) and conveyed the wrong message. I'm working on finding replacements; there are some good ones out there. Just so everyone knows. Jessemv (talk) 03:37, 15 October 2011 (UTC)

I'm making good progress gathering the images and getting permissions. Jessemv (talk) 00:57, 17 October 2011 (UTC)

Permission has been granted by Vijay Pande. My next step is to choose the best image and get it onto the article! Jessemv (talk) 01:51, 17 October 2011 (UTC)

Most of the images I was looking at have been added. There may be better ones out there, but at this point I'm going to go back to focusing on improving the text. Jessemv (talk) 23:52, 19 October 2011 (UTC)

Nonsensical sentence in the introduction?

Can someone clarify what this means?

"The Folding@home project has also successfully simulated protein folding in the 1.5 millisecond range — which is a simulation thousands of times longer than it was previously thought possible to model, and millions of times longer than ever previously achieved."

The source says that they successfully replicated a folding situation that takes a long time in a short period of time.

But the sentence on the page says the simulation of 1.5 milliseconds is 1000x longer than it was previously thought was possible? So they are talking about 1.5ms/1000? This makes no sense. The million times longer makes even less sense.

Jabberwockgee (talk) 02:57, 20 October 2011 (UTC)

Hi Jabberwockgee, the source says that their first goal was to break the microsecond barrier, and the the millisecond barrier is 1000x times harder. I'm pretty sure there are 1000 microseconds in a millisecond, so that makes sense. The "thousands of times longer than it was previously thought possible to model, and millions of times longer than ever previously achieved" statements comes partly from the main site http://folding.stanford.edu/ and http://folding.stanford.edu/English/FAQ-Press#ntoc8 but neither of them say exactly that, which is odd since I was sure it was covered by a source, but now I can't seem to find it. Hmm. Well, it does seem a bit braggish and perhaps untrue now, so I've replaced it with something better that is fully backed up. Thanks for pointing it out. Jessemv (talk) 04:24, 20 October 2011 (UTC)