Message boards :
Number crunching :
Draining My Q's
Message board moderation
Author | Message |
---|---|
Send message Joined: 30 Mar 08 Posts: 50 Credit: 11,593,755 RAC: 0 |
I'm more than willing to continue with MW as my full time project. It is unfortunate that recent events reveal a seriously flawed playing field. I can no longer justify the expense of contributing to MW. The failure of the project to field optimized code for all (emphasis on all) with more than adequate server support is the main reason I will be looking to sign on to another Boinc study. I have given this a lot of thought. My evaluation of current circumstances plus the most recent admin message about inefficiencies and the implied organizational failures that support same leaves me with some very uncomfortable feelings about MW. I am looking for a project that offers a "full pull" for my contributions. I know I offer one in return. All it will take is a level playing field. So simple yet so hard to find. Regards-Voltron Vcore...Vcore...gotta make it sweat to score. |
Send message Joined: 21 Dec 07 Posts: 69 Credit: 7,048,412 RAC: 0 |
I've come to the same conclusion. With a few contributors running an optimised app, the project doesn't appear to need me any more. I was going to hang around until I reached the next big milestone, but it's become pointless. My computers worked hard to get into the top 10, only to be knocked out in a matter of days by one of the optimisers (possibly two of them by the looks of things today). P.S. I am not looking to attack the project admins, nor do I agree with the actions taken by those who have discovered the flaws in the applications' source code. I just feel it's fruitless to continue while my efforts are being overshadowed by people running third party applications. |
Send message Joined: 21 Aug 08 Posts: 625 Credit: 558,425 RAC: 0 |
I've come to the same conclusion. With a few contributors running an optimised app, the project doesn't appear to need me any more. I stopped processing because I was concerned about validity of the results. I'm not a "high RAC" / "Top 10" person though... As for your stated reason (rankings), to each their own I suppose...but as I've pointed out elsewhere, they could achieve the same high RAC by having hundreds of very fast machines, but with them all hidden... I could complain how I don't think it's "fair" that you have 5 active relatively fast machines while I simply don't have money to buy a newer machine. Where does one draw the line on "fairness"? |
Send message Joined: 21 Dec 07 Posts: 69 Credit: 7,048,412 RAC: 0 |
As for your stated reason (rankings), to each their own I suppose...but as I've pointed out elsewhere, they could achieve the same high RAC by having hundreds of very fast machines, but with them all hidden... I could complain how I don't think it's "fair" that you have 5 active relatively fast machines while I simply don't have money to buy a newer machine. Where does one draw the line on "fairness"? I only mention rankings to highlight how insignificant previously significant contributions have become, compared to those machines running the optimised apps. I didn't say anything about fairness. Put it this way... A week ago, Milkyway handed out approx. 3 million credits a day (per BoincStats). Yesterday, the top two crunchers got nearly that much between them and at least one of those two is using third party apps. So even if everyone else disappeared, the project would still be getting about the same amount of work done. Hence "the project doesn't appear to need me any more." Yes, I agree with your comment about validity of results. |
Send message Joined: 21 Aug 08 Posts: 625 Credit: 558,425 RAC: 0 |
As for your stated reason (rankings), to each their own I suppose...but as I've pointed out elsewhere, they could achieve the same high RAC by having hundreds of very fast machines, but with them all hidden... I could complain how I don't think it's "fair" that you have 5 active relatively fast machines while I simply don't have money to buy a newer machine. Where does one draw the line on "fairness"? Ah, but you did...or at least that's what comes across when reviewing other posts about how you wanted to "get to 4 million 'the hard way'"... Bear in mind I am not trying to attack you, but only try to shift this discussion a bit. You have 5 systems active and, as of this post, have a RAC of 20,980. Voltron has 20 systems active with a RAC of 82,336. Their current RAC is at 519,964. All Voltron has to do is add just over 6 times the horsepower. Is that "expensive"? Yep. If I had the money to do that and actually did it, would I also be subject to the scrutiny and the aspersions if I was a silent participant that only posted one message here on the forums for a little bit of startup help for a problem with a new host and had my computers hidden? What about NEZ over at SETI? Since the computers are hidden, you never know how many of them there are. It could be just 20 or so that have some super-duper optimized app. SETI is open-source on their app too...so it is a possibility. The project has allowed 3rd party apps. Just because they are better than what the project has now does not mean that those individuals are doing anything dishonest. If they are and are not telling anyone, shame on them...and delete their credits and results (if invalid). However, if you want to be upset at the project, be upset, but I'll tell you that if I was in Dave or Travis's position, I too would be cautious about such a dramatic difference with someone on the other side telling me that they're better than me, tease and taunt with snippets of code, and that I should get to work and "figure it out" when I had just told them that I didn't write the code and I'm working on another area of the code.
To make sure we're talking about the same thing, I mean that I was concerned about the validity of the application that all the rest of us are using, given the claimed slopiness... |
Send message Joined: 21 Dec 07 Posts: 69 Credit: 7,048,412 RAC: 0 |
Apples and oranges as far as I'm concerned. But if we have to compare it to SETI, they have a direct link to their source code, easy to find, so anybody can download it and (provided they know how), work their magic on it. Plus there are sites (and links to sites) that provide optimised applications. So, getting back to Milkyway, you say "The project has allowed 3rd party apps". Where are the links for the source code and faster third party apps for Milkyway? As for validity, yes, I read it the other way around and no, I don't know the answer to that. |
Send message Joined: 21 Aug 08 Posts: 625 Credit: 558,425 RAC: 0 |
SETI is an established project. This is not. Also you'll note my new thread that I started where I've posted code. I guess it's not obvious to click on the GMLE: Distributed Generic Likelihood Evaluation hyperlink on the home page, then read that page and find the text that says "The package can be downloaded here." which has a hyperlink that starts an HTTP file download, but it didn't take me much time to figure that out... Edit: I'm off to bed... Goodnight... Edit2: You need something to open the .tar file if on Windows. WinRAR is a good choice. 7-Zip might do it as well...not sure. For Linux, you are golden out of the box... |
Send message Joined: 28 Aug 07 Posts: 146 Credit: 10,694,581 RAC: 4,226 |
Where are the links for the source code and faster third party apps for Milkyway? Well, seems it went down under in this thread. Crunch3r posted the link to it there earlier. Seems not everyone who wants it noticed it and is hard to find only by a forum search anyway. Here you can find the source code finally: http://www.cs.rpi.edu/~deselt/ Member of BOINC@Heidelberg and ATA! My BOINCstats |
Send message Joined: 1 Oct 08 Posts: 106 Credit: 24,162,445 RAC: 0 |
However, if you want to be upset at the project, be upset, but I'll tell you that if I was in Dave or Travis's position, I too would be cautious about such a dramatic difference with someone on the other side telling me that they're better than me, tease and taunt with snippets of code, and that I should get to work and "figure it out" when I had just told them that I didn't write the code and I'm working on another area of the code. Travis knows for this for at least half a year. It is not that it took him at surprise. We just wanted to make that fact more public to increase the pressure to change something. That they work on something else is the excuse for half a year. Why do you think Crunch3r got frustrated after working with the project? Another example is Ramsey at home. They've (better: he has) also put out quite bad code, but it is a newer project than MW and he is actually reacting to user comments. Just look how long it took for him to produce a better (still not very good) version. Full two days. And Ramsey@home is not run by a bunch of computer scientists. |
Send message Joined: 7 Jun 08 Posts: 464 Credit: 56,639,936 RAC: 0 |
Not that I'm disagreeing on the 'quality' of the current application code here... But you all seem to be neglecting one detail here. The current application works on all the supported platforms without a lot of farting around for any one concerned, and I'm going to assume they all come from the same code base. IOW's you attach a host and it works without having to fight through a string of compute errors first. Therefore, since the project is getting the output from the hosts they are seeking at this time from their science POV, why should they go and complicate their lives by diddling around with it right now when they may have other more important issues to deal with which aren't apparent to us? The old adage of if it ain't broken (from the functional viewpoint), then don't fix it comes to mind here. Clunky beats offline in my book any day. <edit> Come to think of it, MS built their way to market dominance, using that exact philosophy! ;-) Alinator |
Send message Joined: 1 Oct 08 Posts: 106 Credit: 24,162,445 RAC: 0 |
But you all seem to be neglecting one detail here. The current application works on all the supported platforms without a lot of farting around for any one concerned, and I'm going to assume they all come from the same code base. IOW's you attach a host and it works without having to fight through a string of compute errors first. That's the reason I emphasize that my app is still normal C code that would also compile on all platforms. There is no platform specific stuff I've put in. |
Send message Joined: 29 Dec 07 Posts: 16 Credit: 158,120,935 RAC: 0 |
Yup, I'm on my way outta here shortly also for many of the same reasons I've seen posted here already. No point rehashing. I hope the project admins someday get their act together, it was fun while it lasted. Asta la vista fellow boinc'ers. Catch ya in some other projects. |
Send message Joined: 27 Aug 07 Posts: 915 Credit: 1,503,319 RAC: 0 |
What about NEZ over at SETI? Since the computers are hidden, you never know how many of them there are. It could be just 20 or so that have some super-duper optimized app. SETI is open-source on their app too...so it is a possibility. AFAIK Nez runs the stock app. But if you really want to talk about RAC just look at the #1 @ Einstein. me@rescam.org |
Send message Joined: 9 Jul 08 Posts: 85 Credit: 44,842,651 RAC: 0 |
Yeah, but it's only leading the pack because it's a machine running 4 6-core dunningtons. Not too hard to lead the pack with 24 cores in a single box. ;) |
Send message Joined: 21 Aug 08 Posts: 625 Credit: 558,425 RAC: 0 |
Yeah, but it's only leading the pack because it's a machine running 4 6-core dunningtons. ...as opposed to 24 Coors in a single box... |
Send message Joined: 9 Jul 08 Posts: 85 Credit: 44,842,651 RAC: 0 |
Yeah, but it's only leading the pack because it's a machine running 4 6-core dunningtons. Yeah... that's a virtual guarantee that you will NOT be leading the pack. :D |
Send message Joined: 21 Aug 08 Posts: 625 Credit: 558,425 RAC: 0 |
Yeah, but it's only leading the pack because it's a machine running 4 6-core dunningtons. I think you may be a wee bit wrong there... |
Send message Joined: 27 Aug 07 Posts: 915 Credit: 1,503,319 RAC: 0 |
Yeah, but it's only leading the pack because it's a machine running 4 6-core dunningtons. After drinking the pack one must wee. me@rescam.org |
©2024 Astroinformatics Group