Message boards :
Number crunching :
astropulse not crediting
Message board moderation
Previous · 1 · 2 · 3 · Next
Author | Message |
---|---|
tullio Send message Joined: 9 Apr 04 Posts: 8797 Credit: 2,930,782 RAC: 1 |
In Einstein@home I have switched from a S5R3 optimized app to a S5R4 non optimized app with no trouble. I cannot understand why this was not possible in SETI. Always running Linux. Tullio |
Byron S Goodgame Send message Joined: 16 Jan 06 Posts: 1145 Credit: 3,936,993 RAC: 0 |
Hm, if something designed to cover flaw - is it "good will" or just "cover for own flaw" ? ;) The difference here is this isn't testing that is causing this, and the people that are using the old apps were told they were responsible for keeping it up to date. So I can understand if the project decides to live up to that to end the issue with the two different versions. On the other hand if the project decides they want to keep giving credit to those folks, I respect that as well, as long as at some point something somewhat consistent is obtained again regarding the credits and how tasks are determined to be valid. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
In Einstein@home I have switched from a S5R3 optimized app to a S5R4 non optimized app with no trouble. I cannot understand why this was not possible in SETI. Always running Linux. Welcome to our club, no one can understand, why Einstein@home way of upgrading had been ignored here. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
and the people that are using the old apps were told they were responsible for keeping it up to date. ... NO, NO and NO ! They don't responsible. They just should be honored for participation AND be pleased to upgrade. They don't have any responsibilities ! They did favor to project by participation, they did even bigger favor to take work for installing opt app. Don't forget, we can't tell something to all participants. We can only post messages on forum boards. And not all participants read this forum. And again, they don't have any responsibility to read this forum. If project decides it needs another application for working, it's PROJECT RESPONSIBILITY to do such upgrade in best possible way, it's NOT responsibility of participants (and not responsibility of authors of opt app BTW, cause we can't do anything with berkeley's servers!). |
Byron S Goodgame Send message Joined: 16 Jan 06 Posts: 1145 Credit: 3,936,993 RAC: 0 |
and the people that are using the old apps were told they were responsible for keeping it up to date. ... I see your point and understand your point of view. I appreciate that all the volunteers here give their cpu time and energy, however as with most things, it has to be by the rules of the project to be fair to all and work within the constraints the project has. I wish the transition could have been done without any blurps, but if it took everyone losing some credit to get the validation process back on track, I would think it would be worth it to ensure no valid science continues to be lost. At this point it appears unless a script is run, that could well be what's happening. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14654 Credit: 200,643,578 RAC: 874 |
Actually, I think there is some responsibility for users here. If you use the stock application, and the applications issued and maintained by Berkeley, then the responsibility is Berkeley's. They should manage transitions - as they did when setiathome was replaced by setiathome_enhnaced - and they should manage the transitions better than they have done this time. But if users choose to install an optimised application - or anything which uses the BOINC 'anonymous platform' mechanism - then they are actively blocking Berkeley's (and BOINC's) automatic mechanisms. If they choose that route, then I think they inherit the responsibility for mainaining their own applications. Berkeley's responsibility, which they should have exercised but didn't, would be limited to not sending new work to outdated applications. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Berkeley's responsibility, which they should have exercised but didn't, would be limited to not sending new work to outdated applications. Exactly! And to do upgrade to new opt and or not to do upgrade as well as to continue participation or not to continue - these questions are on participant discretion :) |
Byron S Goodgame Send message Joined: 16 Jan 06 Posts: 1145 Credit: 3,936,993 RAC: 0 |
|
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
Berkeley's responsibility, which they should have exercised but didn't, would be limited to not sending new work to outdated applications. I'm afraid already late.... But don't know... Maybe, if they re-release V5.0 under different name... That way it is still possible..... |
tullio Send message Joined: 9 Apr 04 Posts: 8797 Credit: 2,930,782 RAC: 1 |
OK. I know the safest thing is to use the stock apps, which are automatically updated. But, since the AK_V8 optimized app works so well on MB WUs while the stock AP was extremely slow on my Opteron 1210 running Linux, I tried an optimized AP which I found in a SETI forum post, not in a black market. I am sorry and apologize. Mea culpa. mea culpa. mea maxima culpa. But when do I find an optimized app 5.0 for my Linux box? Tullio |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
OK. I know the safest thing is to use the stock apps, which are automatically updatesd. But, since the AK_V8 optimized app works so well on MB WUs while the stock AP was extremely slow on my Opteron 1210 running Linux, I tried an optimized AP which I found in a SETI forum post, not in a black market. I am sorry and apologize. Mea culpa. mea culpa. mea maxima culpa. But when do I find an optimized app 5.0 for my Linux box? LoL :) Sorry, AFAIK there is no AP V5.0 opt app for linux for this moment. There is one ported from lunatics sources but it still has validation problems and under debugging (again, AFAIK ). I don't follow linux porting process too closely now. My very own proposal - to refrain from AP crunching for some time and do SETI MB or even to participate in SETI CUDA MB testing on beta site. But again, it's my very own opinion. |
Byron S Goodgame Send message Joined: 16 Jan 06 Posts: 1145 Credit: 3,936,993 RAC: 0 |
My very own proposal - to refrain from AP crunching for some time and do SETI MB or even to participate in SETI CUDA MB testing on beta site. Good advice and can probably save some folks some aggravation. Me personally, I'm going to just continue to do AP / MB on main and beta and start the CUDA as soon as I clear the cache on one of my pc's. They'll sort the validation between the different apps in time, I'm not going to let it deter my doing any of them. |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
Berkeley's responsibility, which they should have exercised but didn't, would be limited to not sending new work to outdated applications. Actually, the problem is that the APP_INFO.XML file lets you override whatever application the project wants you to use. You could set up an APP_INFO.XML to process Multibeam with Notepad if you wanted to. It wouldn't get credit, but you could tell BOINC to try. So, they could/should have changed the application name, but didn't think of the optimized applications and the existing APP_INFO.XML files (that they did not provide). It's a big deal here in the forums because the only practical way to find the optimized applications is to read about them here. As of a few months ago, out of roughly 900,000 crunchers, only 4,200 had posted within the last year. Getting bit is bad, but like it or not, the ones who have been are a small minority. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
[quote] Well, hope "small minority" way of thinking will be not extended to another spheres of existance..... :| Each disappointed cruncher from this "small minority" set is real human with real nerves and feelings.... BTW, interesting to know how big this "small minority" share in work done. Cause each PC with optimised app is at least 2 times more valuable for project than the same PC with stock app. And using of opt apps is well correlated with best CPUs using , with 24/7 crunching and with farm of PCs holding. So I think this "small minority" is pretty big share of project power indeed. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14654 Credit: 200,643,578 RAC: 874 |
Berkeley's responsibility, which they should have exercised but didn't, would be limited to not sending new work to outdated applications. The app_info is certainly an aggravation of the problem, but it isn't the whole problem. Did you follow my running debate with mmonroe through most of yesterday, in the Astropulse Errors-Optimized version 5 thread? He was running the stock application, and got bitten by a zero credit (not a pending credit, mind) from another stock application host: a simple 'no reply' timeout and reissue, using fully standard automatic BOINC tools on both hosts. The Berkeley rollout should have handled that contingency too, but it didn't. |
Strider Send message Joined: 12 Oct 99 Posts: 26 Credit: 50,019,051 RAC: 0 |
That's happened to me twice....I do the calc's for 286,000 seconds for a credit of 794.00 and my granted credit is 00.00 That's the last time I will do an astropulse. I'll just delete them when they come in... Chuck |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14654 Credit: 200,643,578 RAC: 874 |
That's happened to me twice....I do the calc's for 286,000 seconds for a credit of 794.00 and my granted credit is 00.00 That's the last time I will do an astropulse. I'll just delete them when they come in... 1) Check whether they stayed at 0.00 - some get full credit when the third wingman checks in. 2) If you really don't want to do them, turn off the supply at source (preferences page) - that way you save Berkeley's, and your own, download bandwidth. |
Raistmer Send message Joined: 16 Jun 01 Posts: 6325 Credit: 106,370,077 RAC: 121 |
That's happened to me twice....I do the calc's for 286,000 seconds for a credit of 794.00 and my granted credit is 00.00 That's the last time I will do an astropulse. I'll just delete them when they come in... Not right way to go. Better listen to Richard and disable AP work in web preferences on website. Much better than aborting already downloaded tasks. |
Richard Haselgrove Send message Joined: 4 Jul 99 Posts: 14654 Credit: 200,643,578 RAC: 874 |
Strider must have been one of the very first to be caught this way: 1061292797 3978228 17 Nov 2008 15:56:25 UTC 25 Nov 2008 15:02:23 UTC Over Success Done 286,418.70 759.74 0.00 His is the first one - 17 Nov 2008 15:56:25 UTC: the 19 Nov 2008 19:38:43 UTC resend got the stock v5.00 Again, no sign of an app_info.xml between the pair of them. |
Sergej O. S. Send message Joined: 29 Oct 08 Posts: 123 Credit: 44,886 RAC: 0 |
Here I have a situation where I was using 5.0, wingman 4.36, and third anonimous wingman 5.0 is working right now. Let's see if all three get credit. Two 5.0 got credits, 4.36 left behind with zero for it state "invalid" what once was "valid". All stock. |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.