Message boards :
Number crunching :
BOINC needs a overhaul
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · Next
Author | Message |
---|---|
Gary Charpentier Send message Joined: 25 Dec 00 Posts: 30640 Credit: 53,134,872 RAC: 32 |
Ah, put a blocker on all new features until all bugs are fixed. Then put a blocker on all new features, until a error reporting system is in place for the public release code. I like it. Let them figure out who writes the pay check! |
Aurora Borealis Send message Joined: 14 Jan 01 Posts: 3075 Credit: 5,631,463 RAC: 0 |
Ah, put a blocker on all new features until all bugs are fixed. I like the idea. Revert back to V5.10.45 until a whole new scheduler is written. It should only take a year or two. I don't use the GPU features anyway. I was fairly happy with the way V5 handled my CPU based projects. Boinc V7.2.42 Win7 i5 3.33G 4GB, GTX470 |
hiamps Send message Joined: 23 May 99 Posts: 4292 Credit: 72,971,319 RAC: 0 |
(off topic) Boy Richard, What a cop out..."And most of the large loads are self-inflicted by people who deliberately push the envelope far beyond what BOINC was designed for" Why don't you let us know here and now what the Load is that is accecptable? Should I7's and one Cuda card be OK as long as they don't go over a GTS 275? Do they get 2 Cuda cards if all they have is a duel core? Should everyone that runs Seti be ordered not to upgrade? I am sorry but that was a lame answer to me! Official Abuser of Boinc Buttons... And no good credit hound! |
Luke Send message Joined: 31 Dec 06 Posts: 2546 Credit: 817,560 RAC: 0 |
I count 14 votes... Luke - Yes Is there anyway the BOINC development team could at least be informed of this thread? So that they know there are groups of users who want change? I've never tried to contact them before. Gizbar summed it up perfectly... BOINC seems to be upgrading for BOINC's sake and not for the community or for the science. BOINC, which literally means Berkeley Open Infrastructure for Network Computing, was designed for distributed computing. I.e. its main priority was the advancement of science and said distributed computing. When it focuses more stability issues, performance issues, and (quite possibly) security issues, I think it's time for a reboot. 52 Aces & Richard, I'll reply or PM you both you later, I've gotta' get to work now. Regards all (and sorry for the short reply), Luke. - Luke. |
Matthew S. McCleary Send message Joined: 9 Sep 99 Posts: 121 Credit: 2,288,242 RAC: 0 |
I'm in favor. I'm just pessimistic about the outcome. But I wish you best of luck. |
Jord Send message Joined: 9 Jun 99 Posts: 15184 Credit: 4,362,181 RAC: 3 |
It also seems that there is a positive, ready, community willing to contribute and discuss many ideas for releases, except it isn't harnessed by the developers:Reporting about it on these forums does not make any sense as the developers HARDLY read these forums."and in my humble view that is terribly inefficient. a) The quoted line was mine. I put that into the development change logs to warn you that if you write your reports on the BOINC Dev forums or any (obscure) project forums, that the developers will not read about it. What good does that do to fixing bugs you have found? Before anyone says, why don't I forward it to them then? I have done so in the past, but it's quite difficult playing middle man if more information is needed. That's why the developers (and I) request you use the email lists. b) Given that such a line gives so many problems, my next choice will be between rewriting it so you do understand what you are asking for, or just completely quit writing change log threads for the development versions. Get them yourselves from the Alpha list if you need to. I was only giving a bit of service. c) As has always been the case, BOINC is Open Source. If you think you can do better, go get the source code and make it better. d) Don't forget that BOINC isn't just a client, it's the back-end that runs on the servers as well. Make sure that after you 'fixed' the client, that you make sure it'll run on the back-end still. e) If that's too much trouble, see if Synecdoche does what you want. If not, tell its developers (if they're still around) about your gripes and what you need/want 'fixed'. Winterknight wrote: "Well I for one think having to keep a computer on waiting for uploads to finish before it can get any downloads is STUPID. It's there to protect against your client crashing and all of your work - preferably - making deadline. Together with the project wide back-off. For what is the use to download work all the time if you can't upload it anyway? All that work trying to upload, and remember at that time it was still trying to upload two results at the same time, failing, then continue with the next pair. All the way through your (long) list of uploads, before starting at the first ones again. Eventually that would completely stop your client, as all it was trying to do was upload all that work you'd accumulated. And in this case, you could still download new work, compute it and add it to the already long list of work impossible to upload. Some people at the time lost their complete queues due to crashing clients, or due to them thinking they'd know better and try to 'fix' things. he said it was put in at the request of another project You're posting on its forums, although Rosetta may have had a slight of hand in it as well. :-) Unable to Upload again Panic Mode On (17) Server problems Panic Mode On (18) Server problems Busy Bytes And from the BOINC Dev email list archives, Optimizing uploads... (archive is free for all to read, no registration necessary). Still, since this "poll" addresses the BOINC developers, I have forwarded the thread to them. |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
Sigh. When the developers state exactly why something was done, there are a few users that refuse to listen. Most of the decisions are based on problems encountered when the code was some other way. 1) The project wide back off was made to speed recovery after an outage. Everyone hitting the servers slows everything down. This change has lowered the recovery time from 2 days to 1 day. In other words, it is better for everyone on average. 2) The rule of preventing downloads when there are > 2 * ncpus results waiting to upload was put in place for a couple of projects where the upload time exceeded the compute time. Not having this in place would have infinite work accumulate on your computer waiting to upload. In other words, there has to be SOME limit. Admittedly 2* is a bit arbitrary. You might make more progress with an argument for a different limit. BOINC WIKI |
52 Aces Send message Joined: 7 Jan 02 Posts: 497 Credit: 14,261,068 RAC: 67 |
52 Aces - No (?) Correct, it's a 'no.' If anyone should be angry about the last outage, it's me ... I ended up tanking North of 340 finished WU's (and harmed 2 others in the process). However, it was 85% my own damn fault in a desperate moment to overcome something that was a split-fault between S@H, Berkeley, and Murphy's Law. Listen, you're all sharp guys and gals. I wouldn't be here, replying, trying, and somedays even sighing if you weren't. But to blame or vent at boinc for this and previous frustrations is (A) misplaced, and (B) unproductive even if it wasn't misplaced. Every software project of merit has a benvolent dictator(BD), possibly with a short hierarchy. Having a BD is even more important in a volunteer (hobby non-revenue) project. It's the only way it works. And in the case of Boinc, that BD is DA. It's "He Decides," not "We Decide." Benevolent Dictators of projects with merit respond to only one thing --> more merit (weighted for features, time, and money). It's not a democracy by vote, or decision by committee. And if you really have merit and really want something done, you do it yourself and contribute it back for review. Heck, if given a choice between those who Pitch-In vs those who Bitch-In, who would you pick? Mere words are free, and worth every penny. [Edit] And for those who recall a post I made awhile ago, yes, it is still the USER community who decides whether something is a bug or not, but it is ultimately the developer who decides whether to do something about it or not. At the end of the day, the source code wins. [Edit#2] DOH!! Benevolent |
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0 |
I'm going to get utterly scathed for saying this... BUT, before you do rip me apart, LISTEN to what I have to say. Anything follows from a false premise, and this thread contains the most selfish, ignorant set of "demands" I have seen in years of participating. First false premise: the problems are caused by code-bloat. Every feature in BOINC exists because it tries to solve some problem. Strip back to the very beginning, and all of those problems will still exist. Some code will need to be added to work on each problem, and very likely, the replacement code will be the same. The second false premise: Code features are added specifically to piss off users. First of all, Mark, you simply aren't that important, and you are no more qualified to decide what is a coding disaster than I am to dispense medical advice. Your self-centered view is that BOINC must make you happy at the expense of everyone else. The project wide backoff will improve recovery after difficulty, if idiots could only leave it alone and watch. Probably the biggest false premise: software development by democratic vote. Who votes? The end users like msattler who are constantly whining because BOINC doesn't work the way he thinks, or the projects that are looking at spending a lot more money just because they have to provide an order of magnitude more capacity to keep stupid people happy. With the possible exception of JM7, no one here is qualified to determine that BOINC needs to be re-written. With 40 years experience, I'm not qualified without a thorough code review -- and that'd take months. I say "possible exception" in John's case because he knows a lot of the code, but he's the only one who could say if he knows enough. Saying the developers should read this forum is just plain arrogant. If they read here, shouldn't they also read Einstein, and CPDN, and every other forum? Doesn't that mean every single thread?? ... and you want a complete re-write??? When can they do that if they have to bend to the whim of the childish forum members who can't be bothered -- reading forum posts can be a full time job. If you don't like it, deal with it, because it is the truth, and no amount of whining is going to fix it. Either way, I'm done. SETI@Home is about finding intelligent life, and 99% of the threads show an appalling lack of clue -- not to mention plain courtesy. P.S. Malevolent? Really?? (every word I can think of to put here would get me banned -- it's tempting, really) |
52 Aces Send message Joined: 7 Jan 02 Posts: 497 Credit: 14,261,068 RAC: 67 |
P.S. Malevolent? Really?? No, not really, dang Latin lapsed me for a moment. The correct phrase (Benevolent Dictator) is however pretty familiar in Open Source circles. |
Odan Send message Joined: 8 May 03 Posts: 91 Credit: 15,331,177 RAC: 0 |
Slightly off topic but I feel I have to share this with you. It is always very tempting during a software development project, especially an indefinitely long one like BOINC that is always trying to keep up with moving goals, to reach a point where someone says "hey, this code is a mess, it is bloated & patched & ugly & inefficient". The following thought is usually "I know, let's start again & do it properly" Several months down the line of "doing it properly" the team usually feels like holding hands and all jumping off a cliff together. They have gradually realised that the new improved version has lots of new bugs & even more frustratingly old & long ago fixed bugs start to surface. I know this, been there, done it, worn out the teashirt & then done it again. I never said I wasn't stupid :) Here is a link to an old hand at this who has put it rather well: http://www.joelonsoftware.com/articles/fog0000000069.html Anyone involved in professional software development or just interested in the process would do well to browse around Joel's amusing & thought provoking website. Don't get me wrong, I'm not saying all is perfect with BOINC & SETI. I just think it is very good & achieves its aims amazingly well. |
OzzFan Send message Joined: 9 Apr 02 Posts: 15691 Credit: 84,761,841 RAC: 28 |
Excellent link, Odan. Loved the read and I think it aptly applies here. |
Odan Send message Joined: 8 May 03 Posts: 91 Credit: 15,331,177 RAC: 0 |
Excellent link, Odan. Loved the read and I think it aptly applies here. Glad you liked it. Do browse around his site when you have time. Much wisdom & humor there. |
Pappa Send message Joined: 9 Jan 00 Posts: 2562 Credit: 12,301,681 RAC: 0 |
It has been a Long Day... Odan, Thank You. Over the Years, I myself have tested many Boinc Core Releases. I know that others that have posted here or have not posted also have tested many of those releases. Over those same years, I am responsible for a bug fix or two (or the introduction of a new bug as a result of a "fix"). There is nothing in the science of programming that ever states that what you are trying to correct will actually correct the problem. The only positive thing is that you are trying. For myself, while I do not agree with the direction that the Deveopers have gone (in all cases). I do agree, they have not stopped working. One of my recent statements into Boinc Alpha was "fix" the Bugs we know and then move ahead. To me that makes sense. Yes I have reported a few that were not reproducable (or reported by YOU). So while it is easy, from where you sit to say "this is broken." Prove and Report it! I and Others have been doing that for a very long time. If this offends you, get over it and do something positive! It does not mean we have to agree... It does mean YOU have to do Something! It only costs you a bit of time. Credits be D@mmed... So Yes, I am a Boinc Alpha Tester. I am a Lunatics Alpha Tester, I am a Seti Beta Tester. Get over it. Regards Please consider a Donation to the Seti Project. |
W-K 666 Send message Joined: 18 May 99 Posts: 19048 Credit: 40,757,560 RAC: 67 |
Sigh. From a Seti perpective. The ordinary user installs and forgets, so one assumes that the cache of tasks is ~0.25days, 6 hours. On a modern cpu the shortest tasks, VHAR's, using the stock application, can be done in under an hour. On Tuesdays the project can be down for the full length of a working day, 8+ hrs. The normal recovery period after the maintenance can be up to 5 or 6 hours. So if the host is on all of this period and has not been able to upload all of the tasks completed, it could have been on for 14 hrs with only 6 hrs of work. This I believe is not how BOINC and the projects were meant to work, i.e. use all the spare cpu cycles when the computer is on. Therefore when setting limits or default settings, the designers need to work out realistic settings, using scenario's from all projects, not just appease the one's that request a change. And if the change requested doesn't fit the majority of projects it must be rejected. I would suggest that the 2*cpu's rule be rejected on these grounds. |
Luke Send message Joined: 31 Dec 06 Posts: 2546 Credit: 817,560 RAC: 0 |
I will politely disagree :). I've had many different problems with BOINC over the past few months. Before I get into details, a good piece of software will function correctly around it's enviroment (i.e. Windows) - programs should work with the host, not the other way round. Blaming Windows is not an answer before I start :). Sometimes I close BOINC down, next time I open it it's stuck in suspend mode and I can't get it out, it gets blocked by UAC in Vista, it can't handle in excess of 4000 tasks well, clicking the icon on W7 seems to open another instance of the program even if one is already running.
Are they quickly forgotten because the documentation is poorly done? ;) I can't remember the post here, but that is what someone said.
It doesn't need my seal of approval, although I would like to know that BOINC is safe, secure and easy (which it is definitely not at times - just look at all the users posting here with problems). Regards, Luke. - Luke. |
Gary Charpentier Send message Joined: 25 Dec 00 Posts: 30640 Credit: 53,134,872 RAC: 32 |
BUG REPORT Boinc uses Alpha a two way e-mail list to report bugs. 1) Alpha is a horrible name for this error reporting list as the user knows he isn't running Alpha software or Beta software, he is running release software and will continue to look in vain for a place to report errors or bugs on release software. 2) Users should not have to have their e-mail boxes stuffed with other error reports to report a bug. They are not developers! 3) Users should not have to be familiar with arcane log switches and know how to edit config files to report bugs. END BUG REPORT SUGGESTION Place a report a bug in a menu item that automatically generates the necessary items for the developers, sends the report and user contact information if more detailed things are needed to find the bug. END SUGGESTION Stop forcing the users to be developers, give them the respect to be just a user. Effective interfaces do not concern the user with the inner workings of the system. That goes for error reporting too! |
Luke Send message Joined: 31 Dec 06 Posts: 2546 Credit: 817,560 RAC: 0 |
First false premise: They are not demands as they are more of a polite suggestion to the dev community. I'm not going to give an ultimatum, and leave if it isn't met, I'll just keep on crunching.
Every program has some form of code bloat. It's how much it contains that is the critical factor here. I know to a few users I have discussed with think that there is code bloat in the BOINC software.
Directed at Mark, no need to reply.
You don't have to be qualified to state a reason or opinion in most cases. People listen to politics all the time, you don't need to be a politician to have a say. We all get to vote don't we?
I implied the BOINC message boards, I thought that was pretty obvious. That's just stupid to make them read every different project forum.
By "fix it" you acknowledge a problem with the BOINC system. The 14 people who propose a rewrite of the software aren't whining, they are expressing they're opinion. ----- With the finances they have and the staff, they do a great job at building the software. Don't get me wrong. I just think they're time could be put to better you by doing a ground up design instead of (bluntly put) bug-ridden minor versions. - Luke. |
dnolan Send message Joined: 30 Aug 01 Posts: 1228 Credit: 47,779,411 RAC: 32 |
without expressing my own opinion, we've heard that 0.007% of the user base feels a change is needed. How do the rest of the Boinc user base feel? I suspect that the majority (more than 50%) of the user base don't even realize there's ever any problem at all. So, rest of the users, how about it? -Dave (no, I don't really expect them to answer... that's the point really...) |
Luke Send message Joined: 31 Dec 06 Posts: 2546 Credit: 817,560 RAC: 0 |
Hi dnolan... Well, by my reckoning, out of the 22 or so people who have expressed their opinion about 15 would like to see an overhaul... 70% or so. That's a very small pool of users though. - Luke. - Luke. |
©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.