Message boards :
Number crunching :
WU Credits
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 . . . 7 · Next
Author | Message |
---|---|
Send message Joined: 12 May 08 Posts: 30 Credit: 1,589,348 RAC: 0 ![]() ![]() |
Hey I'm late to this talk-fest but where's the optimized application so everyone can use it? Otherwise ban it. published by DAVE on 1st of may: http://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=307 |
![]() ![]() Send message Joined: 27 Aug 07 Posts: 46 Credit: 8,529,766 RAC: 0 ![]() ![]() |
Hey I'm late to this talk-fest but where's the optimized application so everyone can use it? Otherwise ban it. The optimized application is not available at the link given in that thread. It is only used by 1 person, who keeps it to himself for the moment to compensate for credits he personally lost in other projects because of interventions by the so-called Credit Cops. BOINC.BE: For Belgians who love the smell of glowing red cpu's in the morning Tutta55's Lair |
Send message Joined: 29 Aug 07 Posts: 486 Credit: 576,548,171 RAC: 0 ![]() ![]() |
Hey I'm late to this talk-fest but where's the optimized application so everyone can use it? Otherwise ban it. I'm pretty sure Those Applications are available Tutta, their just not the Application Crunch3r was using ... :) [edit]Sorry Tutta, after re-reading your Post I see you did say "The optimized application" which I assume means the Optimized Aplication Crunch3r was/is using[/edit] |
![]() Send message Joined: 29 Aug 07 Posts: 12 Credit: 510,143,409 RAC: 2,863 ![]() ![]() ![]() |
|
![]() Send message Joined: 14 Jan 08 Posts: 3 Credit: 641,912 RAC: 0 ![]() ![]() |
I do not think everyone will be satisfied for quite some time. On one hand, I think people should get equal credit for equal work, which is (as far as I can tell) what we have now. The issue is that the optimized app is so much faster than the regular app that you have one person with rediculous amounts of credit. In reality, the project is probably giving out too much credit per WU, but it works for the regular app. The ideal situation would be to develop optimized apps for all processors and reduce the credits per WU, but that is not a reasonable request for those running the project. Also, we apparently are not at a stage where this would be helpful to the project. In the end, the log scale is not a perfect solution, but it is the most reasonable at the moment. After reading Travis' posts, I have to agree with him. P.S. If there is anyone out there who is a good coder and can optimize the app, please do and solve this situation. Until then, I say accept the current situation as this project is still in development and will not be perfect. I will continue to crunch for MW, and I hope everyone else does too. |
![]() ![]() Send message Joined: 22 Nov 07 Posts: 285 Credit: 1,076,786,368 RAC: 0 ![]() ![]() |
This means people who take the time to optimize an app ...... will be optimizing apps and running other projects! Why waste time to optimize an app for a project that will limit you. 2. The average milkyway user doesn't see the credit they've been getting decreased, or have it made meaningless by a couple people running optimized apps getting extreme (like the mountain dew commercials) amounts of credit WAIT! you mean my credit is meaningless? you mean to tell me that I am not going to be able to go out and get that new fancy car with my BOINC credits! 3. We keep the amount of credit our project is awarding in line with what other projects are doing and don't ruffle any feathers or make our project look bad. Cross project parity is a Dave Anderson wet dream. He has created a situation where people are no longer dedicated and crunching for his pet project SETI, so in an effort to bring them back, he has proposed that all projects should bend to his wishes and grant the same as SETI. Another scientific doofus. Come on do it- grant 10,000 credit for every work unit returned,, I know you want to! if it was about the science the optimized code would be openly available so other people with a similar processor could use it and help the project get results faster. As it stands, a couple thousand volunteered hosts are helping the project a whole lot more than a single optimized host; so it's the sciences best interest to keep them happy. Complaining and dropping the project because a host can't get 200+ credit for a few minutes of work doesn't seem to me like science is the issue here. More bla bla bla. Science is not free, neither is it open. Hopefully BOINC can change that, but so far it has not. The history of Crunch3r and his optimized code goes WAY BACK to the early BOINC days, where he got slammed for giving out his code. So, for him not to want to get slammed again and again, he has wisely chosen not to give it out to the masses. You know what they say once bitten twice shy? The rate we're getting results at currently is good enough for what we're trying to do. This is still an alpha project, with good reason. Good, then you wont be missing those that leave because they feel limiting a host for doing valid science faster than anyone else is not scientific. I am going to miss these chats, I really am. |
Send message Joined: 21 Aug 08 Posts: 625 Credit: 558,425 RAC: 0 ![]() ![]() |
Some observations, from an "outsider"... I think some individuals in the other thread are a touch green with envy... If any of you have read my posts elsewhere, you should get the impression that I don't believe in any of the current forms of "cross-project parity", nor do I believe in using credits alone to try to compare rankings among projects because of the differing "standards" used to derive credits (as someone else puts it, the differing "basis" for credits)... This situation is one that breathes life into the current proponents of "cross-project parity". The perceived "unfairness" of someone else getting more than others. I've seen messages here about how it is unfair that he has the application, but nobody else does. There are a few facts to consider though:
As I stated before, someone paid for the tools being used to create the application he used. I was more "gentle" with what I said to the other participants, but I'm going to be a bit more open with you: It is your job as "Project Administrator" to ensure that the app gets integrated into the project as a whole, not his. He was and is under no requirement to distribute, particularly when you all haven't fully officially signed off on the executable's use. Widespread use of that application would likely have a tremendous impact on your project's capability of serving all attached participants due to the increased demand on the server-side infrastructure due to much more rapid processing of tasks. He actually has handled things THE RIGHT WAY, giving you the chance to review things, and thus you should be a bit more tactful in how you deal with this situation. Finally, I think that you need to spend a little more time mentioning that the results submitted by the optimized application seem to be equivalent to the results submitted by the standard ("stock") application, given the prior use of terms like "fishy" that were made before you knew one way or the other... Your statements only fueled the fire that was already burning out of control... |
![]() ![]() Send message Joined: 22 Nov 07 Posts: 285 Credit: 1,076,786,368 RAC: 0 ![]() ![]() |
Hey I'm late to this talk-fest but where's the optimized application so everyone can use it? Otherwise ban it. GO and download the source code, and build your own. And if you decide that you want to share what you have done with everyone else great - I will take copy of it. If you decide not to allow the masses to have access to your work (like Crunc3r has) well then that is up to you. |
![]() ![]() Send message Joined: 22 Nov 07 Posts: 285 Credit: 1,076,786,368 RAC: 0 ![]() ![]() |
Brian, Well said.!!!! But not just the tax code, but the entire free trade and open market. What Travis is doing is more socialistic than scientific. And for those that have envy towards others for their ability's, too bad. Wish I had the knowledge and smarts of Steve Jobs, but I don't. |
Send message Joined: 21 Aug 08 Posts: 625 Credit: 558,425 RAC: 0 ![]() ![]() |
He's only following the trend. It appears to be the "social norm" to try to conform on the level of credits. It's not that I disagree with that general concept. I don't. I actually think it would be nice if a project stands or falls based on its' merit, which is exactly what the cross-project parity thing is supposed to "do". Where I have problems is that the benchmark system and the "flops counting" mechanisms can both be "gamed". Even if there was not intentional tampering, there are enough variations between operating systems and even processes that are running that can cause variances that make me ask: "parity" in relation to who? As for "envy", I am having the "Despair.com" posteritis lately, so this is something for those folks to keep in mind: ![]() ...where those that are envious are attempting to perform the role of the "lawnmower". It is awkward how society seems to wish to punish success... |
![]() ![]() Send message Joined: 12 Nov 07 Posts: 2425 Credit: 524,164 RAC: 0 ![]() ![]() |
WAIT! you mean my credit is meaningless? you mean to tell me that I am not going to be able to go out and get that new fancy car with my BOINC credits! Darn! I was saving up too. Doesn't expecting the unexpected make the unexpected the expected? If it makes sense, DON'T do it. |
Send message Joined: 7 Jun 08 Posts: 464 Credit: 56,639,936 RAC: 0 ![]() ![]() |
I not saying that I don't agree with you in principle when it comes to folks taking open source applications and optimizing them to run more efficiently. According to the GPL they have that right. It's not about 'envy' over someone's coding skills, or having the cake to go out and buy the latest cutting edge technology every 18 months. The question is what's fair under the rules and givens of the game. The problem is you have to be careful when you start making direct analogies to economic and currency systems. The reason is that there is not one fixed basis currency in use today, and the BOINC credit system was designed with that as a given right from the start. The issue is since there is no reference authority to set that basis independently, similar to the way you can set your computer's clock to the reference primary time sources for example, it puts the burden of determining that basis according to the definition of the Cobblestone on every single individual project. If everybody plays to the same set of rules, then it works reasonably well. If not, then credit and everything else derived from it has no meaning. Alinator |
![]() Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 ![]() ![]() |
How am i supposed to do this? i have no more access to his source code than you do. I also am not a code optimizer; so i really don't have the knowledge to do so appropriately -- i also don't have the platforms to do this for every possible processor our users use, and these kind of things are individually tailored to specific processors. Like i said before, optimized apps really aren't a top priority for us right now, so wasting a few months of my time to do this really is not in our best interests.
What was fishy was the performance improvement. I raised the amount of time WUs were being kept on the server to check out the results and make sure the WUs were correct. ![]() |
![]() ![]() Send message Joined: 22 Nov 07 Posts: 285 Credit: 1,076,786,368 RAC: 0 ![]() ![]() |
I raised the amount of time WUs were being kept on the server to check out the results and make sure the WUs were correct. And what have you found so far? Again, if the result is valid, then so is the credit. And limiting the credit for a valid result is not being fair and equal. How am i supposed to do this? i have no more access to his source code than you do. It is your project is it not? Ask him for the code, and review it. But better yet, just do what you say you are already doing, review the returned result for validity. If it is valid, then again, the credit is valid. Period. |
![]() Send message Joined: 30 Aug 07 Posts: 2046 Credit: 26,480 RAC: 0 ![]() ![]() |
I raised the amount of time WUs were being kept on the server to check out the results and make sure the WUs were correct. It's not fair. It's a compromise between what's good for the project, our average user, and what code optimizers want. I'm not going to have the project be giving out a couple hundred credit for a couple minutes of work, sorry. ![]() |
Send message Joined: 21 Aug 08 Posts: 625 Credit: 558,425 RAC: 0 ![]() ![]() |
From what I understand, the concepts behind the code were made available to you and to Dave. That included what to look at and what generally to do. If you want the code, then you could ask, couldn't you? However, if you're going to allow independent optimization at a later date anyway, will you also be asking those people for their code? If not, is that "fair"? Is punishing him publicly "fair"? If you think it is, then the follow-up question would be: Is it "professional"? I also am not a code optimizer; so i really don't have the knowledge to do so appropriately -- i also don't have the platforms to do this for every possible processor our users use, and these kind of things are individually tailored to specific processors. Like i said before, optimized apps really aren't a top priority for us right now, so wasting a few months of my time to do this really is not in our best interests. Since you're not an optimizer, then you might also be making assumptions that may not be true. There may be some code that is specific to specific processors, but generally they will be for classes of processors that support a particular feature set. Examples: SSE, SSE2. Providing that type of support in your standard application would provide the project a scientific boost for all processors at least a Pentium III or AthlonXP or newer for SSE, Pentium 4 / Athlon64 and newer for SSE2. SSE4.1 tweaks are all nice and stuff, and I'm sure he probably did do things like that for his specific architecture, but the things that are specific are likely in a much lower proportion to things that are generic across multiple processors. If optimized applications are not a top priority, and you're going to take action against their use, then I would suggest that you secure your source code and only allow official applications until such time as you are able to devote more time to it...
You should've not said anything. You should've done the actions behind the scenes, raising the amount of time like you did, then you could've come back out and said that the results were valid without throwing fuel on the fire. |
Send message Joined: 17 May 08 Posts: 16 Credit: 528,507 RAC: 0 ![]() ![]() |
I raised the amount of time WUs were being kept on the server to check out the results and make sure the WUs were correct. I agree 100%, FULL credit should be given to valid results, no matter how fast it is getting done, by whomever and whatever. Period. |
Send message Joined: 7 Jun 08 Posts: 464 Credit: 56,639,936 RAC: 0 ![]() ![]() |
I raised the amount of time WUs were being kept on the server to check out the results and make sure the WUs were correct. OK, so I ask that either you or zeitgeistmovie to define 'Full Credit'. Alinator |
Send message Joined: 21 Aug 08 Posts: 625 Credit: 558,425 RAC: 0 ![]() ![]() |
I raised the amount of time WUs were being kept on the server to check out the results and make sure the WUs were correct. As I said, the progressive system you came up with will never fly with the hardcore "parity" people because, well, it won't bring "parity". Eventually you're going to have to make a decision on whether or not to lower the basis for your credit determination, as it already is "too high", per various individuals. Just for kicks, since I signed up here so I could be able to post, when I get a moment I'll attach my AMD system and see how much credit I get from it... If it is more than about 18/hour, then you're already in trouble with the Credit Cops... |
Send message Joined: 21 Aug 08 Posts: 625 Credit: 558,425 RAC: 0 ![]() ![]() |
I raised the amount of time WUs were being kept on the server to check out the results and make sure the WUs were correct. I can't get a good read on timings for comparable tasks due to the quorum of 1 and that they are purged so fast, but my view is that the credit awarded should be equal to the slower application, just that the faster application got it done quicker. Thus, if the slower ("stock") application was to get 200 credits for 4 hours of work, if his faster application turned back in the same results in 10 minutes, then he should get 200 credits, no more, no less... |
©2024 Astroinformatics Group