Dear Mr Allen,
23 Apr 2006 22:11:21 UTC
Topic 191127
(moderation:
at first I have to say that you should be proud to run such a spectacular project, involving thousands of volunteers providing Teraflops of Home PC calc power.
Oh yes, the projects' success is the work of the volunteers, and their only benefit from participating is (besides satisfaction to support science research) to compete in that little contest counting their cobblestones.
As for many others, it was a great thing for me to view the ranking every morning how our team was prospering and how my personal position changed overnight. Of course, there are guys running BOINC on huge computer centers I can hardly compete with. But that's life!
Then, within the last weeks, a horrific enemy showed up. Yes, it's your 638 Dual Core Opteron weapon I'm talking about. And I'm really astonished about that: Is it for your personal vanity that you're competing with home computer users?
Hey, you're supposed to crunch on every machine available to you, for sure. Because it's your job to do so. I wouldn't have said a word if under your name the iBook and the PowerMac showed up. Also those two boxes would not make me complain. But the entire cluster under your name? Come on, you must be kidding us ;)
It should be a competition among volunteer crunchers. Or are the team and user stats something i would in german call "Schwanzvergleich", proving you have the biggest one? ;)))
And finally, a serious word to say: Throughout the last week I found the project servers down for several reasons. Meanwhile, there's at least one hour downtime to expect per day. How comes you're running 600+ computers under your name and the project servers suffer from hardware issues almost daily? How about spending a few of your boxes to increase system stability. It would not cost your first place in the team ranking, for sure.
Ok. I wonder if I will ever receive serious answers on that one. Happy crunching to all of you out there.
--lox
Dear Mr Allen,
)
The news explains the outage over the last week. A bad circuit breaker, which was finally replaced. There is nothing that can be done about those, and the outages they cause.
I think it is excellent that
)
I think it is excellent that the leader of this project is demonstrating his committment to completing the work along with all us volunteers.
The fact that the credits are counted towards his name or a group name like the UW-Madison CAE is unimportant.
If this were a competition, you should quit now while you are behind. Based on your RAC, you are not going to reach #1.
Unlike other projects, Mr. Allen and his University are supporting the project with cpu cycles and power just like you and I am.
Your slander does not deserve a responce from him or any of the project team.
Of course it is nice that he
)
Of course it is nice that he is actually participating. But why does he get credit points for that? He could participate without competing with the rest of the participants, couldn't he? It just doesn't sound fair to me.
RE: Of course it is nice
)
Fair? It's not like he's stealing a prize or something; there is no such thing. Methinks you folks need to chill out your over-competitive blood just a mite. :-)
microcraft
"The arc of history is long, but it bends toward justice" - MLK
RE: Fair? It's not like
)
The ranking *is* the prize. I suppose for most participants the ranking is the key reason to participate. Otherwise there wouldn't be links to the "top teams", the "top participant" and the "top computers". There wouldn't be various websites providing ranking and additional statistics. And people wouldn't carry a stats display around in their signature.
RE: at first I have to say
)
Bruce and all the people who made E@H work on this end are proud. The volunteers who made E@H a success should be proud as well.
We are aware that the volunteers are the heart of E@H.
We try our best to give the volunteers a positive experience as a small repayment for their contributions. Even though we don't always succeed it is not because the we do not appreciate all that the volunteers have done for us.
I don't know what to say about the cobblestone issue. I will leave that to Bruce.
For those who are interested Bruce is running E@H on the brand new NEMO beowulf cluster. NEMO is not stable yet and E@H is being used to burn the cluster in. In the future NEMO will primarily be used to look for gravitational waves from inspirals.
The E@H downtime has nothing to do with Bruce running E@H on the nemo cluster.
The recent downtime resulted from:
* some flaky hardware in the database server (possible bad mobo?)
Last Friday I moved all the E@H servers from the old room on the 3rd floor of the UWM physics department to a to a new data center on the 2nd floor. The new data center was primarily built for the NEMO cluster but there is plenty of room for E@H as well. There is a giant UPS for the entire room and there is tons of cooling so I am very happy with our new home.
Today the database server finally had a complete failure and so we are now running off the backup server.
We are not completely out of the woods yet but I have new servers here and more hardware on the way that should make E@H fast and reliable on this end. I will have the computers I need but it will take some time on my part to get everything working.
There is your answer. I know we don't always communicate the best but you can rest assured that we are working hard behind the scenes.
Once again thank you to everyone who has contributed.
P.S. I appologize if I got too defensive.
Just my 2 cents on this
)
Just my 2 cents on this 'sensitive' issue of credits and rankings.
David, thanks for the inside info. I don't see why you should be on the defensive here. In my opinion the project is going on steady, and I hope it will yield important results.
For me E@H is the only way I could contribute a little to 'hard' science, being a philosophy major and a translator at present. Credit is a relative measure of what crunching power and time every one of us is able and willing to contrubute. It's good only while it's fun. I simply stopped looking at my credit since I downloaded one of akosf's optimized clients. It's good that my PC returns the results faster. Why should I compare my credit with people who are using the standard client?
I'm glad that we have the opportunity to help. Isn't it enough? :-)
From what I gather is, that
)
From what I gather is, that it seems to be unfair for home users being matched up with the cluster computers. I do understand the frustration of my fellow team members having put in a heck of a lot of machines, hard work and a lot of determination only for a cluster to relegate them to second in no time.
While on a personal level I don't mind, this does not help remove frustration levels. I guess, the core essence of the message is, please find a way of for home users to better compare each other's contribution. Having a cluster in the list of course takes a way the motivation for those groups of participants not only doing it for the cause of science but also because it is a bit of a sport and therefore fun.
Imagine in car racing all are driving with a two litre engine and suddenly one shows up with 8-litre turbo. Of course, everybody would call this unfair because it is crystal clear that's a competition. Here the thought of "competition" is somewhat fostered too, for those who would like to compete in terms of overall contribution.
I experience the buzz in my team when the other teams where closing in on us and we somehow pulled some more computing power out of the hat and then happy posting all over the place! Of course, that is great motivation and that is also going on in all the other team forums. Please project team, understand that there is a part of participants seeing this also as a kind of sporting event which generates fun and excitement. For some this is the primary reason of why to stay project bound for longer than the odd month - but for year after year.
You may now turn around and say that this belief or mental frameowrk is wrong, that it is only for science, but then you would take away the fun for a good portion of participants. The same effect has the provision of a supercomputer that outguns pretty much everything, just think of the motorsport example above.
All I am asking from the project board is to look for a way into how to make this event for those who also see this as a kind of sport exciting and fair once again. Maybe by introducing a flag signalling the computer class (i.e. home, work, cluster, etc.) of that there is a new ranking class for team containing supercomputers.
To my understanding of course Bruce should participate and he should lead by a shining example. But I guess, that this massive display of computer power does not help some eager determined and prepared user groups. Therefore, I kindly ask to review measures of how to put the fun back into E@H for those users. C'mon, we are dealing here with astrophysice, so that shouldn't be too difficult to achieve, is it? ;-)
Regards
:
your thoughts - the ways :: the knowledge - your space
:
RE: understanding of course
)
As I said in another post, it does not seem to me that those Dual Core AMD Opteron are doing a great job. My poor old 400 MHz Pentium II completed a WU in 77k s using Albert 4.51 while an Opteron finished in 16k s using Albert 4.40. Not a great show.
Tullio
Am I looking at another
)
Am I looking at another contribution vs credit thingy...?
Cause if it does, maybe I'm starting to regret joining this project.