Looking for DC project.

[flyted]

Have done some DC in the past, but had often found them to be somewhat unethical in that they had ties to private companies or that you couldn't find out how much management was paid. Sometimes it just plain appeared we were working for them for free and not doing much good. I know for some they don't care about any of these things and thats fine, for them. But when I donate, I want to know the whole story. WCG and FAD had financial ties that were at best very suspect so they are off the list. FAH seemed more like a free service for Stanford with little results for the amount CPU time they got, yet I did a lot of work for them before I gave up on it. Would really like to see some in-depth coverage like a 20/20, 60 minutes or the like to spotlight the good DC projects and weed out the bad. Too much potential for abuse here.

Not looking for an argument on the subject as the facts are readily available to those who dig deep about some of the projects , just want ideas and any input from others who also screened their project to make sure their donation of lots of CPU time was being used for a good purpose. I think thats a fair request.

 

[RaySun2Be]

So what exactly is your criteria for choosing a DC project?

Almost all, if not all, use our PCs for free to help them with their project.

And as far as the validity of F@H, see the link here: http://forums.anandtech.com/messageview...atid=39&threadid=1950979&enterthread=y where that very thing was discussed.

Too much potential for abuse here.

What kind of abuse are you talking about? Using our PCs for free? No expected results for hours/years of processing time?

There are many projects running under the BOINC system now, maybe you could find a project or two that meet all of your criteria.

Most of us are willing to donate our time, PCs, electricity to causes we enjoy, we think will help fight disease, or better understand scientific information or just for the sheer fun of stats and participation.

I don't know of any that have gotten "burned" by a project other than loss of stats, etc.

Nor do I think any of the "abuses" you are concerned about have come about, or even warrant a 60 minutes expose. I just don't think it's there.

 

[flyted]

DC is no different than a charity. Who doesn't want what they donate to charity to go to good use? Thats a reasonable expectation that everyone can understand. Are all charities good? If you donate 30 minutes a day of a Celeron processor then there is nothing much to lose. If you have dedicated machines, its a different story and there are costs involved. I have 1 machine that I bought new about 2 years ago that was solely for DC projects. For some, these DC projects are like a religion and will defend them at all costs and yet have done no research into them. Maybe they give to charity and don't look into them either. Thats fine, its their choice. I choose to know what I'm donating to, thats my choice. There is a very high potential for abuse here if you look, and no it has little to do with no instant results. DC doesn't give instant results. It has to do with how the program is structured, how long the program will last, how practical the program is, how much people are being paid to run the program, ties to private industry,etc. My post was not to debate whether anyone should DC because of points or the fun of playing on the computer, nor have I told anyone what they should do. My interest is in finding a health-oriented project that is well run and accountable. After my post I finally found the first attempt I've seen to present a realistic view of DC projects that didn't show them all as perfect. This is the type of info I'm looking for, hope the link works.


http://www.hyper.net/dc-howto.html

 

[caferace]

Originally posted by: flytedNot looking for an argument on the subject as the facts are readily available to those who dig deep about some of the projects...

Please. That's a troll if I ever heard one. "I'm right, so don't bother to argue and if you do you're ignorant".

Have you ever wondered if people have different motivations, or even interpret data (deep or not) differently from yourself? Is your world really this black and white?

-jim

 

[RaySun2Be]

I guess I still object to the term "abuse". The article you link to makes some great points, and thinks to think about when choosing a DC project, but I don't think any of the ones I'm familiar with would even remotely be considered abusive. They all have stated objectives, goals, targets, etc.

There is potential for "abuse", but many of the DC participants question the project teams, double check what's going on, so the potential for true "abuse" is fairly low, and would be uncovered and announced quickly.

As always, it all boils down to a personal preference.

 

[flyted]

Caferace- Sorry to see that you are not capable of an intelligent discussion, or of reading the link I posted.

RaySun2be- As you can see, some become very infantile at the idea that their project may not be what they want to believe it is and are incapable of even reading the link I found to get the bigger picture. That link was the only one I found that offered any comparison and it seems the author did a lot of research for it. The term abuse is relative I guess as no one is sending in money, but if someone found out the project they had supported and done a lot of work for had wasted their time or was poorly managed, can't imagine them being happy. Its no different than giving to a charity. They are rated for the same reason. Who sends a check to every "charity" request they get in the mail? Nobody. So why wouldn't anyone want to check out a DC project as in-depth as they could to make sure its worthy of their time and effort? Thanks for at least making the effort to discuss the issue.

 

[Insidious]

HiYa Flyted

I am a little confused as to the thrust of this thread.... Looking for a project - Like the title says, or looking to explain why you have chosen not to participate.

Either emphasis is totally appropriate here and I have appreciated hearing your perspectives. (They are very similar to some of my own)

With most of the same concerns you have expressed, I chose Rosetta@Home. I liked the free sharing of their results to any group that wanted to use them. I also appreciated what appeared to be a very sincere concern from the project developers about the worth of what their donors were providing to them. (I don't feel taken advantage of... at all)

Anyway, you might want to give Rosetta a look.

(I also have a long history with F@H and recently re-started contributing to that project. That decision was based on the team spirit exhibited here at AnandTech among the F@H folks more than the Stanford project itself.)

 

[networkman]

Interesting thread.

How's that saying go... "Just because I'm paranoid doesn't mean there aren't people out there trying to get me."

Truly. I've done a bunch of different DC projects during my time with the TeAm. Some of have been non-profit, some have had a possible carrot at the end in the form of a cash prize(albeit probably with worse odds then winning the lottery), and others have probably been for someone's profit. And I'll be honest - I frankly don't care! More often than not, it's been all about the stats. All about keeping the TeAm in the top spot or near the top anyway. Sometimes, it wasn't even enough to be crunching the units with my own machines - yes, I even picked up SETI work units off of eBay! :shocked: Ask the old-timers they'll confirm I was(and still am) nuts enough to do it.

I've taken the time to get to know every project I've been involved with - no regrets here with a single one of them! So flyted you just go ahead and look for whatever DC project strikes your fancy and agrees with your principles. The rest of us will continue to crunch what interests us and so the TeAm will continue on in glory.

 

[flyted]

Insidious- I can't remember the name now of a DC project I did a couple years ago about animal species that eventually lost funding, but I was in the top 25 there and ran several machines full time. I crunched for TA F@H and was at one time in the top 100 there (leconc) before I stopped, so I would have to disagree with your comment about me looking to explain why I'm not participating. When I find one I like, my participation of total cpu time is more than the vast majority. I have a dedicated machine I built for DC and to run 100% of time and a few other machines part time. Someone who donates 30 minutes a day of Celeron time would have no interest in this thread because the issues of resources and cost important to me wouldn't affect them in the same way. They should still care about not wasting the time they do give as it could all be important in the right project. Very few seem even open to the idea that any DC project could be run les than perfect. Common sense says thats just not true. Again look at any charity, see any difference in how effective they are or how efficiently they use their resources? For some reason DC'ers don't want to look at that, they just want to believe. I haven't asked anyone to change or stop doing anything, I've just asked if anyone has decided to take a closer look. The link did exactly that and I thought maybe, just maybe, a few others looked for themselves also.
I have looked at Rosetta and so far it looks good. Being a new project, how friendly is the interface?

Networkman- Not sure how to take your last paragraph, but like I mentioned to Insidious I never asked you or anyone to justify their project or why they crunch. I'm looking for people who do care about what they are donating to and also cared enough to look past the surface of a project. You can crunch for any reason you want and for any project you want, in fact you don't even need a reason. I will however assume due to the quote at the bottom of your post that you didn't mean your comments as I interpreted them.

 

[Insidious]

I personally like rant threads. (Authored a few myelf.) I'm not sure I ever bitched out someone who was actually trying to see my point of view though. ... Interesting twist

 

[BlackMountainCow]

[info]

I tink the project you're talking about was Lifemapper, right?

[/info]

 

[Philippart]

Rosetta's probably the friendliest boinc project available.

you get many updates directly from the project scientists and the community is very helpful.

 

[flyted]

Thanks, Lifemapper was the one, great project. Hadn't heard the name in so long I forgot it.

 

[flyted]

Philippart-- Give me the quick overiew of Rosetta from th operaritonal user standpoint. Is it like most others in that you download the software, then get assigned wu's, then just back and forth again? Ive read the info page and seems like its kinda new and maybe still needing some software upgrades, how you found that to be true?

 

[Ken g6]

Rosetta is a BOINC project. Once you download BOINC and attach to Rosetta, it will auto-update the Rosetta client if it needs to.

 

[BlackMountainCow]

I can give you an overview as well; I've been doing daily Rosetta stats for the TeAm for quite a while now and own 150,00 creds.

Rosetta is sort of new - in comparison to SETI classic - but for a BOINC project, it's one of the older ones. The software upgrades you mentioned will always be with Rosetta. Because of the nature of the project, the researched proteins vary from task to task. So, from time to time, the Rosey devs update the main application in oder to incoorporate new methods or new science apps. The last time, about a week ago, they implemented Alzheimer's desease algorithms. So, the updates are necessary and are part of an ongoing developement.

Rosetta now uses a fixed credit system, that really gives credit according to the science you did, and not according to benchmark numbers, which are "fakeable". Together with Einstein@Home, Rosetta has one of the fairest credit system in the BOINC world. See here for details.

Rosey also has a dedicated beta test programe, called RAPLH. Before new code, algorithms, WUs or what have go live on the main Rosetta, they'll be tested by the RALPH guys for errors or glitches. That works pretty good - most of the time. Of course, you'll always have a few probs now and then. We're all humans, aren't we?

As far as crunching is concerned, it's pretty simple. Download BOINC, attach to Rosetta, create account, get work, crunch, send, get new work, .... the regular stuff.

As Philippart already mentioned, the devs are pretty good for Rosetta. They're often around on the message boards, like to get feedback, ask their dedicators for their input and the overall tone on the Rosey boards is really friendly - left aside one or two of the never-do-well people or trolls.



To me, Rosetta is the most worthy BOINC project at the moment, both for their goals and their attitude towards flow of information, availability of sources and openess of results.

 

[flyted]

BlackMountain Cow--Thanks for the detailed report. I'll continue looking into Rosetta. Since you mentioned LM, did you participate in that one also?

 

[BlackMountainCow]

Yip!

I finished 13,000 species in 270 days, 16 hours, 27 minutes and 7 seconds, ranking 207 out of 3839 participants. :evil:


Wonder why I know that so well? 'Cause there's a selfmade Lifemapper certificate hanging right in front of me :laugh:

 

[flyted]

BlackMountainCow-- Good deal on the LM stats. Too bad they went belly-up. I had talked to one of the developers on the phone and knew they had lost their funding and stopped a couple months before they ran out of money. Really liked the project. I've decided to give Rosetta a try. I read many posts on their forum but need to find out about running several computers there and potential problems, know of any? I'm doing a test run now(leconc) at reduced cpu time to test for any issues?

 

[BlackMountainCow]

The only problem I'd know of right now is that there are a few ( and I mean very few ) WUs which will get stuck at 1% progress and hang there for like 20 hours. Rosey has an integrated "watchdog", which will kill a WU like that after some 24 hours, but still, it's wasted crunching day, if that happens. But as I said, it's a very rare phenomenon. Oh wait, one more. Sometimes, if you pull up the gfx window or if the screensaver pops in, depending on your gfx card and chipset (ATI/Nvida), the current WU might crash.

But let me put this into perspective. During my last 3,000 creds I had one WU crash on me. That's all. With that WU, I lost 60 minutes of work. As you see, Rosey is pretty stable, but not perfect. Nothing is.


I don't see a problem with running on multiple PCs. I do the same; 5 machines. Just set them all up the same and you're good to go.

 

[flyted]

Ok thanks. I also looked through the Rosetta forums to find out what settings are best for the program but didn't see much on it. I don't care about the points, just what works best for the program, any ideas there? Also does it run better on AMD or Intel proc's and how much ram does it need? I configure my machines to be effecient as possible for the program, suprised I didn't see anything about this in the forums. I'm going to be having more questions as I get all set up, think its better to throw them out here or on the Rosetta site?

 

[BlackMountainCow]

There's no diff I'd no of between AMD and Intel as far as performance is concerned, i. e. no special compiler flags were used.

RAM is a different story. Generally, the more the better. But from what I can see, the current WUs have a memory footprint of max 150MB on my machines. So I guess a 512MB system will do just fine. But I also run Rosetta on my trusty PIII-600 with 192 MB RAM and Win2k and that works fine as well.

Will you run Rosetta exclusively on your PCs? If not, and you combine it with other BOINC projects: there's an option in the general BOINC preferences on the Rosey site, that say "leave app in memory". You should set that one to YES, as it will boost performance and minimize restarting problems when the tasks switch.