Message boards :
Number crunching :
EVERYONE PLEASE READ WHAT HAPPENED TO CRUNCHER3D
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · 6 · Next
Author | Message |
---|---|
Misfit Send message Joined: 21 Jun 01 Posts: 21804 Credit: 2,815,091 RAC: 0 |
Do you have anything of substance to add? Points to refute? I'm all ears (eyes). Oh I didn't misunderstand you at all. I know youre talking about the client. I'm sure you noticed I was absent from all the "credit" threads. That's because I don't care if I get a little credit or a lot. I just want to burn thru as many units as I can in the fastest amount of time. However, this also means that I didn't know about what you said, or what Crunch3r said, or what anyone else said. I didn't know about "multipliers" or their values (and still don't). What I was saying IMO is we should be concentrating on better and faster apps and let the BOINC people worry about the credit as they did originally. me@rescam.org |
W-K 666 Send message Joined: 18 May 99 Posts: 19362 Credit: 40,757,560 RAC: 67 |
Do you have anything of substance to add? Points to refute? I'm all ears (eyes). It's not a 0.01 adjustment that is causing the problem it is a multiplier of 7 (seven) instead of the official 3.25 (or maybe 3.35). |
Misfit Send message Joined: 21 Jun 01 Posts: 21804 Credit: 2,815,091 RAC: 0 |
It's not a 0.01 adjustment that is causing the problem it is a multiplier of 7 (seven) instead of the official 3.25 (or maybe 3.35). I took that from Tony's post of 3.35 and 3.36. But then I removed that line after reading your post. You were just too quick for me. :) Now if SETI gave me cash for my credits.. or even if my credits could be used at the new SETI store then I'd care about credits. Until then it's all about the speed for me. :) me@rescam.org |
zoom3+1=4 Send message Joined: 30 Nov 03 Posts: 66284 Credit: 55,293,173 RAC: 49 |
This really pisses me off! I'm thinking about dropping my support for SETI. I've been crunching since 1999 and crap like this has to stop. Too many great people have left SETI due to some childish ignorant people that need to be terminated from the project. People like Crunch3r and trux have kept the spark alive in a lot of people that otherwise would have gone elsewhere. There is no doubt that the science is the main objective, however you have to understand that competition is what drives the heavy hitters out there. When teams were started they introduced the competition aspect of the project. Teams are for one thing and that competition, it's a fact. Everyone tweaks their systems to get the most out of them as they can and a part of that tweaking was optimizing their systems to use their CPU's to their fullest. Crunch3r I hope you will return later and I hope you get the apologies you deserve. I've suspended communications on My PCs for one week over Crunch3r's treatment and the lack of attention to problems that have been mentioned already, After that I'll open up My Net access again. Boycott SetiBoinc - For 1 week only! So Yeah I'm with You. Savoir-Faire is everywhere! The T1 Trust, T1 Class 4-4-4-4 #5550, America's First HST |
Administrator Send message Joined: 26 Jan 06 Posts: 43 Credit: 13,801 RAC: 0 |
Hard to actually gauge the level of upsettedness here. Upset people talk a lot and talk a lot more, so it makes it seem like a bigger issue than it is. Allow me to summarize what I've gathered are the things people are upset about because of Enhanced: 1) People who were running Optimized before, aren't performing as high above the default app anymore (because the default app is much better). They liked their high RACs and are upset that they are lower. 2)The new science app is a little buggy and hangs without progression sometimes. No one seems to care about the science of this, any opinion I've heard has only screamed "Why am I not getting credits for this time? This is horrible. I demand someone fix it." 3) Work units are too long. 4) Work units can vary in size. 5) People running SETI@Home are disrespectful people who put out a bad update and don't care about their users. They don't care about bugs. They are ignoring everyone's concerns and marching on belligerantly. 6) The method of calculating credit has changed slightly. 7) Even for people running standard apps/clients both before and after, there is a big gap in credits earned. The gap is unfavourable, never favourable. 8) Credit between SETI and other projects is not even. And here are what I can gather is the official responce to those people being upset: 1) They acknowledge there will be a drop in credit. The credit gains from optimized clients were temporary. Enhanced was never hidden, and widely announced. To anyone who understood how optimized apps and calibrating clients got you "extra" credit, this was predictable from a long time ago. It's a bit of a long issue, and, to the best of my knowledge I've explained it with an analogy elsewhere, if you're curious: http://setiathome.berkeley.edu/forum_thread.php?id=30918&nowrap=true#319204 I also offered up several solutions, and why they're far worse than the problems, and left it open for anyone to suggest anything else. No one did. It has been acknowledged, and decided that in terms of this particular complaint, no action will be taken regarding credit. 2) The buggy app is not acceptable. It was not expected. They are spending much of their time getting this fixed ASAP. While low credit given for hung work units is fair (credit is given for results, not effort, and results are stalled too), but very undesirable (hanging work units aren't getting results for them, credits for users, and is making everyone upset). They ask people be patient until they figure this out (soon), and/or help by giving them the information they need to pin it down. 3) Work units were designed to be longer. This has many benefits, the most significant being less burden on the servers which users were complaining about going down regularily due to being maxed out for workloads. This is as much of a solution to those complaints as can be implemented from a simple rearranging perspective. The drawbacks are mild in comparison: some of the slowest computers crunching infrequently won't be usefull, and some users want to see many smaller results per day rather than fewer big results. Can't keep everyone happy, but this is by far a better ballance than they had before (also means they're spending time on science, not fixing crashing servers). This will not be changed. 4) Work units were designed to be variable in size (computation time). They *tend* to be similiar, and fall within a certain range, (they're not off by a factor of 500, they're still pretty close). That is why they did not set up a system to award flat credit per work unit. Variable work units squeezes every bit of results they can out of their input, so it's good for science. If calibration between longer and shorter work units is off, they may recalibrate. They will be watching it for the next while. 5) Enhanced came out for several reasons: A) Because it is more efficient, it has rougly tripled the amount of use they can get out of any given chip. Thereby, it's as if they just tripled their userbase. This is a massive deal. B) Because it squeezes more science out of the same input (I think? It does peform a better search from a given signal, right?). C) To address a common complaint that SETI never does any updates and is not responsive to its vocal userbase. D) To demonstrate the link between fundraising and results, and thank the contributors. The funding drive turned in $250,000 (and they need $750,000), and they wanted to show "this means something, here is something new we can do because of that". When people care and contribute, the science moves forward. 6) The new method of calculating credit is more fair. It properly rewards and more accurately calibrates than the old system. They're monitoring feedback and seeing if complaints are valid, precise to this issue and not some other issue, and are keeping an eye on things. Things seem to be working fine, and this is tentatively declared a non-issue. 7) Taking aside all issues of optimized or non-standard components to anything, there will be slight variations depending on the processor/work unit, but they should be within 10%. The situation is being monitored, and calibration will occur if it should be, but otherwise is as predicted and fine. 8) Credit is supposed to be close to even between BOINC projects. That is how BOINC was designed and one of the conditions of the project. BOINC is not a SETI-only creation, it recieves funding for being a larger system. SETI would need to lower its credit if it is unfairly rewarding its users. Possibly, it could be brought to someone's attention that another project is assigning credit unfairly high. Users were invited to investigate and report back. In several days, very very very very very few people have bothered to (2 or 3). The situation is being monitored and when they have enough data collected, they will make their decision to act or not. --- A common theme I'm seeing everywhere is people are yelling and complaining and making demands and threats, but that's what they want to do. Many people are impatient. For every 100 people that yell, make threats and complaints... when given an opportunity to suggest what should be done, or contribute to solving the problem by giving feedback, 1 person actually does. It's really hard to isolate someone giving credable feedback on a problem from someone just barking about "less credit" and "unfairness" and "being ignored". It's also hard to tell what people are upset about. I think that most of the people who complain are 90% upset that their Optimized credit has mostly vanished, and 10% everything else. But, since them complaining about their optimized RACs changing can be humiliating, because all it does is put light on them making poor assumptions or looking like whiners, they reach out for whatever other little legitimate reason they can, and then completely disproportionately pretend to be upset about that. I suspect the vaste majority of people that are complaining about bugs or fairness or calibration or people at S@H ignoring them, don't really give a rat's ass about those issues. They just want to vent and pout. They're so attached to their RACs, they see it as their "property", and see any drops in it as someone stealing from them. Regardless of why it drops. ... There are some legitimate concerns, I think. And, I'm not saying everyone is acting childish. There are some people who have reason to not like how some things are being done. Just, nowhere near the amount of people that are complaining. If I was involved in S@H, I would probably just ignore the forums entirely. We're mostly useless, and they're actually trying to solve problems and address concerns in their spare time, rather than run damage control for non-issues or resolved issues. Look at how many people here are "Going on strike" and demanding appologies over speculation over something they're not even sure about. "Oh, Crunch3r got so upset he left. I don't know why, but I'M LEAVING!" I've seen almost no one outline what they're upset about. I really had to dig to find things people were upset about, legitimate or not. There's a lot of noise, not a lot of signal. If you want to be listened to, make yourselves worth listening to. If anything, someone making threats would make me want to hold off addressing their concerns.. because then what happens next time? Do you give in every time someone makes a stink, and give in more the more upset they get? It sends the message "This is the way to get things done, yell, complain, and make demands." If you're making a demand, you could be either unreasonably upset, or reasonably upset. The usefull part of either of those, is resonability and how reasonable you are being. Not how upset you are being. Bring an issue to the table reasonably, it will get addressed. If you don't like how it is addressed, address that. If it's not going how you want, choose to leave. But if you bring it to the table with complaints and how it's all about you and all the things you're going to do if things don't go your way and how you're going to force them to act... people are going to question your judgement about the issue and how serious an issue it is because if your response wasn't reasonable, what're the odds your thought-process was? People complained about the S@H people not addressing their concerns. So S@H came here and tried to explain a few things. More flame wars erupted, so.. off they go to actually fix problems, since civilized discussion here is a waste of breath. Possibly the reason S@H doesn't participate much is because of the behavior of the membership here. They might as well just continue ignoring us until they have some issues already resolved, because explaining how they're working on it is only getting met with idiocy. The only message people have sent S@H is "We are too worthless to talk to. Don't keep us in the loop. We'll complain about you ignoring us either way." |
EclipseHA Send message Joined: 28 Jul 99 Posts: 1018 Credit: 530,719 RAC: 0 |
Two things which must be considered: 1) RAC is a meaningless number 2) With enhanced, the credits are finally fair between Windows and non-windows (Linux). Windows has been "overclaiming" since Seti/Boinc went live, and now the playing field has been leveled. Most people here are on Windows, so yes, you're finally claiming close to what would be claimed on the same HW, but running Linux. Be happy that you've been able to "overclaim" all this time. |
Reverend Scott E. Lee Send message Joined: 7 Apr 06 Posts: 1 Credit: 4,974 RAC: 0 |
Okay, here's the deal Ffolkes (old movie reference, look it up). Get over it already. I'm sick of seeing this all over the forums. I'm just a piddly little number cruncher that wants to do WORK. I thought an optimized science client would make me more efficient, but if it's going to be this much of a hassle, I'll just continue to plod along with the stock version. Now, for those of you that care SO MUCH about credit? Is there some kind of contest going on that I wasn't aware of? Is there some kind of reward for having the most credit? Are prizes going to be awarded? Can it be pro-rated for people like me that still use a single box with a 1.7Ghz P4 and 512MB of RAM? Will Bob Barker be handing out the car keys personally? Are you KIDDING ME with this stuff? Get over it and get back to doing the WORK. That's the whole point of BOINC and S@H, isn't it? The rest is just crap. -The Right Reverend Scott E. Lee. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13842 Credit: 208,696,464 RAC: 304 |
Allow me to summarize what I've gathered are the things people are upset about because of Enhanced:.... You've got a lot to learn Administrator. Talking sense around here is just not acceptable behaviour & is likely to get you into very serious trouble. Cease & desist before things get ugly! Grant Darwin NT |
Pepperammi Send message Joined: 3 Apr 99 Posts: 200 Credit: 737,775 RAC: 0 |
Okay, here's the deal Ffolkes (old movie reference, look it up). But what if it was you good name that was trodden into the dirt like is bieng done. Wouldn't you like to stand and defend yourself? 'Right Reverend' wouldn't you like to think there are people out there that are good enought to wont to help defend you too? Just a thought None of us would be happy if we where at the recieving end. The rest about credit I agree. Load of crap. |
Beach Bum Send message Joined: 13 Nov 05 Posts: 178 Credit: 611,717 RAC: 0 |
Get off the linux wagon woody, its a meaningless 5 % of computers in the world, maybe its not windows overclaiming, but linux underclaiming. And thats giving it a moderate percentage. My personal specs on linux from all my website stats, shows it as a meaningless 1.4% on traffic to my sites. Opps I went off topic, setting clock to see how long it takes before this one gets deleted. Oh and yes before you ask I do run some linux boxes as well as windows. Beach Bums Current Stats: Come Join us at Hawaiian Beach Bums |
Beach Bum Send message Joined: 13 Nov 05 Posts: 178 Credit: 611,717 RAC: 0 |
Are you KIDDING ME with this stuff? I did go back to work. I dumped SETI and loaded projects that actaully do some good for the world. SETI is not going to find anything. They are searching for noise on the premise that the aliens are using the same tech. So keep looking if you like. Personally I will spend my money on a better idea nowdays. The whole project needs a good shaking up before a lot of the people that left will come back. There is a problem with the app, people tried to point it out, it was ignored. Now we have WU's looping and hanging, this thing was not ready for the general crunch yet. We have Mods that decide to censor/deleted/hide threads because they don't like what is said, or it is said against their buddies. Yet their buddies can break the posting rules and it stays up. I personally have had it. As many other have. I have detached most of my units, with the rest to follow. I am also going to be talking to the team on a pull out of resources to other projects. Just as SETI USA is also looking at doing. As I believe SETI germany is also talking about. How many is it going to take leaving the project before things really are truely looked at, not band aided. Just my 2 cents. Oh and to the higher ups in SETI, when multiple systems in a farm (running standard app) drop from a return RAC rate of 100-110 each to a return RAC of 20 - 35. There is something wrong. Thats not even talking about a large number of units not updateing since the new app, running loops, and hanging. So its not really about the drop in RAC. Its more about the release of a flawed app, one you knew was flawed but pushed anyway. Has the serious drop in RAC pissed a lot of people off, yes. Why, because for many that is their bragging rights to how much science they crush for the project. I wish nobody here ill will, but my machines will not be back to SETI. Best regards and good luck in your search. Beach Bums Current Stats: Come Join us at Hawaiian Beach Bums |
RePhLeX Send message Joined: 3 Dec 05 Posts: 76 Credit: 128,962 RAC: 0 |
Another excellent response Beach Bum. Let the boycotting commence!.. |
mikey Send message Joined: 17 Dec 99 Posts: 4215 Credit: 3,474,603 RAC: 0 |
Jim - there was sarcazm in what I wrote - I just got one question - why was it removed? Because Tony accused Crunch3r to be a cheater while not contacting with him (his address is all over his website) on the issue that was in this thread or because people - IMHO it was right thing to do (but it is only my opinion) - attacked Tony afterwards? The thread was removed because Tony asked for it to be removed. Tony has now started a new thread and is absically saying the same thing only phrased a bit differently. Seems Tony has a problem with the versions that Crunch3r was using, something about Beta versions being mixed in with released versions. I did not see that Tony had a problem with Crunch3r as a whole, although he does think the optimized versions should be outlawed. He is entitled to his opinion and to have his own threads removed if he chooses. The thread removal, he says in the new thread, is because the old thread started out badly, by him, and got worse, instead of provoking the discussion he had hoped for. |
Kim Vater Send message Joined: 27 May 99 Posts: 227 Credit: 22,743,307 RAC: 0 |
Allow me to boil this "Crunch3r fuzz" down with my '2 cent metaphor': An employer needs a new worker that can operate 3 different machines (at any time) in his workshop. (Let's call these 3 machines for SSE, SSE2 and SSE3). There's only 2 applicants to the job - so the imployer decides to give them both a 14 day trial period (to see how they perform) - before hiring one of them. After the 14 day trial, it turns up that the one worker can only perform with an average of 50% effeciency on all 3 machines - while the other worker is working with near 100% effeciency on all 3 machines. Now - which worker do YOU think got the job ?? PS. A normal sence of logic (IQ) will sure help you to choose the right worker here ;-) My 2 cents Kiva Greetings from Norway Crunch3er & AK-V8 Inside |
HachPi Send message Joined: 2 Aug 99 Posts: 481 Credit: 21,807,425 RAC: 21 |
Thanks to Crunch3r for all the work he has done for free for this community. Shame on those who are trying to pull him through the dirt without a shadow of any proof... Greetings from Belgium, HP ;-)) |
Michael Send message Joined: 21 Aug 99 Posts: 4608 Credit: 7,427,891 RAC: 18 |
The thread removal, he says in the new thread, is because the old thread started out badly, by him, and got worse, instead of provoking the discussion he had hoped for. Sir I do not agree...I think he got the response he was looking for. |
SargeD@SETI.USA Send message Joined: 24 Nov 02 Posts: 957 Credit: 3,848,754 RAC: 0 |
Hard to actually gauge the level of upsettedness here. Upset people talk a lot and talk a lot more, so it makes it seem like a bigger issue than it is. I was awfully tempted to quote your whole message. I would encourage you to read through some of those threads again. Most of the time you will find that concerns were voiced in a reasonable manner, only to be belittled or maligned by the few who think they run the message boards. It was their drowning voices and flames that led to the eruptions you see. I think if you look real hard, you will even see an appeal by me to the "real" powers that be asking them to intercede. Nothing came from them until I sent an e-mail to the Chief Scientist for the project and only then did Eric post. But he did nothing to quell the flaming that we were receiving from the "unofficial" powers on the message boards. So please, do not give partial facts to support what you claim to see and do not bend what was said to match your own preconceived notions of what happened. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13842 Credit: 208,696,464 RAC: 304 |
But he did nothing to quell the flaming that we were receiving from the "unofficial" powers on the message boards. 1 It's not his place to do so. 2 It wouldn't have made any difference if he did unless he said exactly what people wanted to hear. As there are such greatly differing opinions, that wouldn't be possible. 2 You & your supporters were never being flamed, although some of the nosie being produced by some of you (including yourself) was getting close to it. So please, do not give partial facts to support what you claim to see and do not bend what was said to match your own preconceived notions of what happened. Please practise what you preach. EDIT- fixed up bbcode. Grant Darwin NT |
jamin Send message Joined: 21 Mar 06 Posts: 65 Credit: 100,008 RAC: 0 |
Thanks - took a long time to get the answer :( The only thing I was touched in the deleted thread was that he went out and told the world that he is upset of using the beta application on the regular Seti instead of informing the person responsible for this in some more "individual" manner. It's a little bit childlish - you know the problem, know the person you should contact and yet you go out and tell "the world" that you're upset. As I understand Tony apologized for this but it was too late: 1.This caused all the flaming 2.This caused Crunch3r offended 3.and all the posts in this thread The damage has been done - and then the thread disappeared - that caused more and more flaming and accusations - because some have read the thread and some didn't have a chance to :) IMHO the thread should be left to continue or to "die" in natural way... And I got my right to my opinion ;) Why are you starting the thread you want later to be removed? [edit] I got pretty sick while reading this and the "Crunch3r conspiracy" - I think I refrain from posting for a little while ;) just to calm down ;))) |
Francesco Forti Send message Joined: 24 May 00 Posts: 334 Credit: 204,421,005 RAC: 15 |
Here I dont' agree. I have one host (on 16) able to work a 4.18 (optimized) in 2~3 hours (a 3.0 GHz dual core) that now is 24~32 hours with enhanced (optimized). Other host works fine, even if slower, but this is not a problem (was the scope or enhanced). The problem (for me) is that given an average loss of 30% in RAC (that is not a problem for me) I see that speeder hosts have a 50% loss and 1 a lot more, while slow host maybe hav onlsy 10~20% loss. Now this "slow" 3ghz dual core it is producing SAME result than old AMD 2200+ with 1 CPU!!!! (same enhanced/oprimized). Now I have reverted this "slow" (!!) 3.0Ghz to original application and it's working on 13~15 hours per WU. If this method for you is "fair" .... (°_°) I wonder why I should use newer host (that use less energy) if their productivity is so bad. By my opinion the new method does not respect the real power of some host. I don't know why, of corse. I only know that previously it was better, not for HIGH RAC but for the propoortional contribute of each host to the job. I think that if you need more details, you can ask me. Best regards, Francesco |
©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.