WCG, R@H, or Something Else?

To crunch or not to crunch....


  • Total voters
    47
Status
Not open for further replies.

AtomicMoose

[H]F Junkie
Joined
Sep 26, 2001
Messages
11,803
Well, it seems the team was pretty well split down the middle on what to do and I found that interesting. In the pursuit of a democratic process, let's see what the team wants to do.

There is a lot of testing going on by different members. Marty and King (and maybe others) are testing WCG. Myself and a handful of others are testing R@H separately.

If you are interested in R@H testing, go here: Download and Instructions and here: Join Team [H]ard|OCP

If you are interested in World Community Grid (WCG) testing, go here: Register and Install
I don't think an official team has been created as of yet.

Also, please share your thoughts on each project and vote in the poll.

And by request...

Here is a project summary for R@H: http://boinc.bakerlab.org/rosetta/rah_research.php
and one for WCG: http://www.worldcommunitygrid.org/projects_showcase/viewResearch.do
 
I am all for a new project and being a 2 project team, but since i have no intention of switching from F@H, i just select the third option
 
I voted for WCG, I like the idea of it supporting multiple projects and users being able to choose which one if not all. I plan on bringing up some older boxes with it if that's what we decide on.
 
Is there any info on each of these projects that we might be able to get a hold of... Maybe put a summary of what each project does in the first post?



 
I voted for WCG, I like the idea of it supporting multiple projects and users being able to choose which one if not all. I plan on bringing up some older boxes with it if that's what we decide on.
Basically this is what BIONC enables you to do. This is not a WCG only function.

Is there any info on each of these projects that we might be able to get a hold of... Maybe put a summary of what each project does in the first post?
I'll see what I can get for you. :)
 
I have a few boxes folding under WCG, they are to old for F@H so I will keep them there. There was already a [H]ardocp team created so I am folding under it with the user name Klockej.
 
Basically this is what BIONC enables you to do. This is not a WCG only function.

But if we decide on R@H then we can only get team points with that project right, not for all of the BIONC projects, correct?
 
I will try to look into the projects more later, but for now, which seems to be more of a fire and forget program, similar to F@H service mode?
 
But if we decide on R@H then we can only get team points with that project right, not for all of the BIONC projects, correct?
R@H runs on BOINC. For instance...

RH1.png


or this

RH2.png


is my using BOINC for just R@H. If you take a look in Marty's thread, the client is the same, but he is attached to different projects.

To be honest, I need to look into the WCG project a bit more as I am not 100% clear on what their main goal is...especially since they start you off with 5 projects. :confused:
 
I will try to look into the projects more later, but for now, which seems to be more of a fire and forget program, similar to F@H service mode?
Neither are reminiscent of the F@H console client to me.

But, I haven't done much tweaking with them yet....more just testing different OS's and machine specs to see how they score.
 
Are these projects more "older computer" friendly than F@H?

Do you think it's worth putting older machines on one of these other projects rather than having them on F@H or is it just more of a personal choice?
 
R@H runs on BOINC. For instance...

http://img.photobucket.com/albums/v711/atomicmoose/Various/RH1.png

or this

http://img.photobucket.com/albums/v711/atomicmoose/Various/RH2.png

is my using BOINC for just R@H. If you take a look in Marty's thread, the client is the same, but he is attached to different projects.

To be honest, I need to look into the WCG project a bit more as I am not 100% clear on what their main goal is...especially since they start you off with 5 projects. :confused:

But how does it compute individual and team points, is it points just for R@H or is it points for BOINC? If I decide I want to support another project that uses BOINC, will my points/WUs/hours count towards our overall team score or will I be off on my own because I'm not doing R@H?
 
Some heavy reading I'll have to do later... I did catch this fairly quick at the R@H site

"with new DNA cleavage specificity"

I'll hold my vote till I read up some more but I'm definitely cool with dedicating some resources to figuring out ways to improve cleavage. :D

I do think it's a bit unfair of a lot of the members to keep saying that a machine is too slow for F@H though. Just because it's not earning you monster amounts of points does not mean that it's not making a good contribution to the cause. I've got several machines running F@H earning me well under 50 PPD - one is even crunching a whopping 57 point WU. They are still producing results that help.

I am however not objecting to the idea of spreading some of mine or anyones else's resources around to more than one project though. Just saying that even if it is a slow old box, it's still helping.
 
Doesn't anyone else think that it would be a good idea to get the results of some of this testing before the UD guys decide on a new project?

Just for the hell of it, I messed with the BOINC and UD client on WCG over the weekend. For a non SMP box, the UD client seemed to be fine. For SMP boxen, the BOINC client started up two instances automatically to use both of my cores. This is something I would consider important in the long run as more and more boxen become multicore.

I didn't notice slowdown at all with either client (then again I shouldn't with my system specs) but all I was doing was normal browsing and light load stuff. I didn't try playing any games or anything like that while I had it running. Then again I was busy with other things.

Also, this was under WinXP. I have not tried the Linux client out. I might do so in the future but with other things going on, I doubt I will get to it before someone else does.

I don't remember exactly which project it was working on at the time but it was either the cancer or genome project and it looked like it was tearing through the work pretty fast. I'm guessing I completed a couple work units and they were both done in about an hour. That would have been two cores with each spitting a WU out in a little over an hour on the system in my sig.

 
I like the BOINC client a lot. I could see myself switching from F@H just because of the client. The whole thing where it runs as many instances as needed automatically, and how it lets you run as a service, but still open the GUI for management.

I voted for R@H, and I am already on the team.
 
I voted World Community Grid because I have a friend with MD and this is the first project I've seen that is working on it.
 
I haven't voted yet because I still need to get a feel for things.
 
Voted R@H as well since I already have 1 machine running (not returned the first result yet since it has stalled). However, WCG is also a very interesting project so either one would be good for us with a slight advantage to WCG due to the ability to pick your preferred projects and the similarity to UD from what I've gathered.
 
Am I missing where is says when this poll ends?
 
After reading up a bit.. I like the idea of WCG contributing to multiple medical research projects that all aim to further humanity's understanding of human biology and medicine.
R@H seems like a worthy cause also, so I guess it comes down to personal preference...

On a side note, older boxen do make a contribution, BUT the environmental impact ratio to the amount of science they do is horrendous and they should be decommissioned and replaced with more power-efficient/powerful newer boxen.. just my opinion.
 
Am I missing where is says when this poll ends?
Nope, it's open ended for now.
But how does it compute individual and team points, is it points just for R@H or is it points for BOINC? If I decide I want to support another project that uses BOINC, will my points/WUs/hours count towards our overall team score or will I be off on my own because I'm not doing R@H?
Points are assigned to each project far as I can tell.

I like R@H and the BOINC client in general, it's pretty slick.
My exact thoughts. :)
 
WCG seems like it would give the users more options over subprojects as part of the larger project... which would mean that any team member can decide what they would like their resources to go to...

my $0.02

i might participate in the selected project due to recent frustrations with stanford
 
Some heavy reading I'll have to do later... I did catch this fairly quick at the R@H site

"with new DNA cleavage specificity"

I'll hold my vote till I read up some more but I'm definitely cool with dedicating some resources to figuring out ways to improve cleavage. :D

I do think it's a bit unfair of a lot of the members to keep saying that a machine is too slow for F@H though. Just because it's not earning you monster amounts of points does not mean that it's not making a good contribution to the cause. I've got several machines running F@H earning me well under 50 PPD - one is even crunching a whopping 57 point WU. They are still producing results that help.

I am however not objecting to the idea of spreading some of mine or anyones else's resources around to more than one project though. Just saying that even if it is a slow old box, it's still helping.

I have several "older" pc's getting about 7ppd... (PPro 200) but in R@H they hold thier own quite well. Yes, they can make contributions to F@H but it seems like they can make a BIGGER contribution on a project where the WU's aren't as intense. No matter what all my boxen will be doing something, I just want to optimze what they are doing.

F@H has geared thier WU's towards faster CPU's with larger cache. This seems to be a trend that will increase with them. That's fine and it's where all my newer rigs go. But for the older stuff, I just feel, based on the points system anyway, that they do more realative productivity on something like BOINC...
 
After reading up a bit.. I like the idea of WCG contributing to multiple medical research projects that all aim to further humanity's understanding of human biology and medicine.
R@H seems like a worthy cause also, so I guess it comes down to personal preference...

On a side note, older boxen do make a contribution, BUT the environmental impact ratio to the amount of science they do is horrendous and they should be decommissioned and replaced with more power-efficient/powerful newer boxen.. just my opinion.

Environmental impact is an amusing concept. It's quite often never thought through. Honestly, is it more environmental to throw out a system or keep using it? What happens when it hits a landfil. Sure you could "recycle" it, but even then what proof do you have it's actually being recycled and what is the environmental impact of that?

My AC demands for my house are so huge that an extra few kwh/month for the older PC's are in my opinion (okay belief) far less of an impact on the enviroment especially when you factor in the good they are doing on folding projects. Even more so on projects where they can make a bigger contribution.

Also, most of my older boxes are doing time as servers and I can't afford to replace them with something new, nor since they are doing just fine in thier jobs, is there a need to.
 
WCG seems like it would give the users more options over subprojects as part of the larger project... which would mean that any team member can decide what they would like their resources to go to...

my $0.02

i might participate in the selected project due to recent frustrations with stanford

I have the R@H client installed but as far as I can tell it's just the same as the WCG client. I attached it to the Rosetta project. But I can attach to other projects as well right? How does this differ from WCG? I assume there must be some differences at least in terms of how and where my points are tracked. Or must I dowload the BOINC client from WCG to attach to projects other than Rosetta? It seems like it's the same client.

More related to your post, I'm just curious since people talk about this in other threads, but what exactly is it that standord is doing that is frustrating? I'm not trying to be judgemental, just inquisitive. I install the clients and let 'em run... so... I'm not attached to anything politically...
 
Have we made a decision as to when we are going to close the "voting"?



 
Can't you run R@H under the WCG umbrella? Would WGC be limited to the 5x current sub projects/what ever subs come along? IE couldn't I join WCG and otp out of all the default projects but the AIDS stuff, or something like that?
 
Can't you run R@H under the WCG umbrella? Would WGC be limited to the 5x current sub projects/what ever subs come along? IE couldn't I join WCG and otp out of all the default projects but the AIDS stuff, or something like that?


As far as I konw, yes you can.
 
Surely more than 38 people on the [H] team have an opinion ???? Is there a mod-available function that would allow a PM or email to all active UD'rs and active FAH'rs (who, in my opinion should be included) -- just to advise them that opinions are being taken by this poll ?

 
Surely more than 38 people on the [H] team have an opinion ???? Is there a mod-available function that would allow a PM or email to all active UD'rs and active FAH'rs (who, in my opinion should be included) -- just to advise them that opinions are being taken by this poll ?


Historically on the moron of the month polls and such there has been ~ 60 voters total on the high end average from my recollection.
 
Is there a mod-available function that would allow a PM or email to all active UD'rs and active FAH'rs (who, in my opinion should be included) -- just to advise them that opinions are being taken by this poll ?

Unfortunately, there isn't anything quite so wonderful to be had out there. If so, you can bet we'd be using it. ;)
 
I haven't voted simply because I feel there should be a second project for the sake of options, and to spread the team's cpu processing love to more research, but to me it doesn't make a difference to what project we support. All the scientific projects are worthwhile, and are advancing the knowledge in the field they are studying.

Also, I don't see the side of the argument that we need another project because F@H isn't worthwhile on "older" computers. Your computer is still going to be spitting out the same number of cycles on any of these projects. Just because you are running some project that looks like you are getting more work units in -- or more -- points doesn't mean that your computer is doing more work. F@H is optimized well with SSE options and memory management, so if you have a computer capable of running that then yes you will produce more. I don't know if R@H or WCG uses and is advanced in utilizing every last bit of your processor like F@H, but if they don't then I would feel that I'd be better off supporting F@H instead of a less optimized program. I know that F@H is advanced, I have no idea about the others, so I'd be hesitant to move my computers over to them if less of their cycles were being optimally utilized.

I'm also a person that if very conscience of the environment, and I don't see the benefit in running a computer that is very old < 1ghz, or any of those old Celerons whatever their speed, because their output doesn't offset the electricity they are using. Simple as that.

I say we pick the project that has been around the longest, and is most organized, and is easy to pull stats from. It should also have a consistent stats system that cannot be fudged by rogue clients. I don't like stats; I feel they can be helpful in bringing in new members, but if they are only folding for the points they are doing it for the wrong reasons. The stats shouldn't be the determiner of the output of the computer, or the worthwhileness of having the computer do X project. In the case of F@H it's still important that people process the standard wus even if they only give 100ppd. You are still contributing. Would you feel better to go to another distributed project that gives you 900ppd? You computer isn't magically running at nine times the speed it did on the other project. It's the same with the SMP vs. standard wus, so OK Stanford is weighting them more heavily, but again you computer hasn't gotten faster, you are still crunch the same number of bits/bytes whatever. You are still contributing the same amount to science.

I think the best thing to do would be for all you people with old computers go out to your community and donate them to families that are less fortunate, and then with the money that would be going to powering those now donated computers you take it and donate it to a cause of your choosing. That gets rid of the issue of old computers. Now we need to pick the project that is friendly to those people that can't connect to F@H servers for whatever reason, and has been around a long time, and isn't getting paid for the results from the projects, and that the results of the research is open to all.

I've said enough for now, and I'm still not voting.

 
I've said enough for now, and I'm still not voting.

So by all indications from your post, you DO think that there should be another option available other than FAH...(please correct me if I am not extracting the right thoughts from your post). I understand that you have no preference for what that other option should be. Now to me , that would specifiy choice #3 (Replace UD with another project. Please post which.) and you posted your ideas or lack of preference in your post.... You have an opinion connected to one of the poll choices, but won't vote...... I don't get it... Unless your underlying intention is to marginalize the poll.

Lack of participation in this poll will render it useless.

 
Yes I do think there should be another. I just don't have a preference. Sorry. So I can't list a project.

I don't have the time to be out researching projects for those traits that I find desirable. Can anyone else list how long R@H and WCG has been around, and how stable they are, if there points system is stable, and what the main research areas are. In a nice simple clean format, instead of tell us to go look at a link.

Also, how does WCG handle the points? Are they seperated out by the projects, or are they all grouped together? I don't want it becoming a nightmare were we tell new recruits to go download WCG and then we have to direct them on how to change to different projects, and disable others.
 
I don't have the time to be out researching projects for those traits that I find desirable. Can anyone else list how long R@H and WCG has been around, and how stable they are, if there points system is stable, and what the main research areas are. In a nice simple clean format, instead of tell us to go look at a link.

Also, how does WCG handle the points? Are they seperated out by the projects, or are they all grouped together? I don't want it becoming a nightmare were we tell new recruits to go download WCG and then we have to direct them on how to change to different projects, and disable others.

I have to agree that these are important considerations to take into account, anyone have any of these answers? I've been away from my primary boxen for the last week, so I've not had any time to put some of this stuff through its paces.
 
Status
Not open for further replies.
Back
Top