|
61)
Message boards :
BOINC Manager :
How best to deploy and manage BOINC to several computers?
(Message 2353)
Posted 31 Dec 2005 by Paul D. Buck Post: Not that I know of... To control all the boxes once you have things installed and running is to run BOINC View. The only time I have to VNC to my remote boxes is when I want to work on the disk drives or to install new version, etc. |
|
62)
Message boards :
BOINC Manager :
I want to ask.... (I am a newbie)
(Message 2352)
Posted 31 Dec 2005 by Paul D. Buck Post: I run,usually, 5 projects on all my computers (the exact 5 varies by computer), work is downloaded from one or more of the 5 and work processing commences. On the one CPU systems one work unit at a time is processed, but after an hour or so it changes to anther work unit from another project. On multiple CPU systems, the only difference is as Bill Said, more than one work unit is in process at a time. On a 4 CPU system I have 4 work units being processed, but they may be all from one project, or a mix of 4 projects. Over time, work is done for all of the projects based on my preferences and other settings. |
|
63)
Message boards :
BOINC client :
How should I assign maximum disk space usage?
(Message 2346)
Posted 30 Dec 2005 by Paul D. Buck Post: There has been a need to update the CLI part with changes, but, as with the rest I have not gotten there. This really should refer to that ... |
|
64)
Message boards :
BOINC client :
How should I assign maximum disk space usage?
(Message 2339)
Posted 30 Dec 2005 by Paul D. Buck Post: Sorry, the cpdn FAQ hasn't been updated for a long while. 600M was for early models, in the days before BOINC. Well, the FAQ in the Wiki is pretty up to date, though I just "tweaked" the disk space numbers to these higher values. |
|
65)
Message boards :
BOINC client :
MD5 checksum error on several projects.
(Message 2317)
Posted 29 Dec 2005 by Paul D. Buck Post: The difficulty is that there are so many things that can cause these errors. Including feuding ISPs. IF he is having problems with SETI@Home he is best off asking on the SAH forums (in general) because the widest readership is there. The down side is I have not seen the network smart people around the last few days ... but, you can try some of the techniques in the Netowork trooublshooting guide in the Wiki ... |
|
66)
Message boards :
BOINC Manager :
Connection Problems
(Message 2308)
Posted 29 Dec 2005 by Paul D. Buck Post: Jon, Well, we *DO* understand. It is just that there is little that we can do about it except help as best as we can. So... There is no doubt that this is a serious weakness in BOINC, and to be honest, I don't know of any good cure that works for everyone. All I can say is that there is no good time for a project shutdown and migration. If it was not this set of problems it would have been another. All I can say is that in general once you get it up and running it is not a big problem. And, there are times when it does seem harder than it needs to be. But, in all honesty, BOINC is getting better. So, when you are ready we will be here. If you will take the time to explain the problem I am sure "we" can help ... most likely it will be Bill or Tony helping ... so the royal "we" ... |
|
67)
Message boards :
BOINC Manager :
Problems with AMD 64X2
(Message 2305)
Posted 28 Dec 2005 by Paul D. Buck Post: Another way to see if it is an overheating problem is to just stop BOINC for an hour or so, then restart it and see if it works. Also, start the computer. SHUT DOWN BOINC; run for 4 hours, then start BOINC, does this problem pattern occur? Can you get the CPU temps from the BIOS? This is a strange one. But, since the computer is newest, I would be leaning to a heat problem (heatsink too small, not properly installed, etc.) or perhaps buggy BIOS. Have you looked to see if there is an updated BIOS? |
|
68)
Message boards :
BOINC Manager :
Exiting with Data Left in Memory
(Message 2295)
Posted 28 Dec 2005 by Paul D. Buck Post: One more point, there are programs where the simulation *HAS* to be run from end to end and there is no possibility that they can be checkpointed. I am not sure that is case with Rosetta@Home yet, but, for example, CPDN can be very sensitive to stopping and restarting the models. There was some work (Folding@Home?, which is in Alpha test) that did not or does not check point at all. and the work runs for a day or so ... So, yes, it can be inattention on the project's part, or, just not done yet, not practical to do, etc. Like many problems in system design there may not be a simple answer, contrary to some. And I have not "met" a project staffer that wants to have any more waste than is unavoidable. For all these reasons, *I* recommend 1) Don't run projects in testing 2) If the project is doing something you don't like, vote with your feet. |
|
69)
Message boards :
BOINC Manager :
Manage more than one client from one GUI
(Message 2294)
Posted 28 Dec 2005 by Paul D. Buck Post: If you have linux only I have heard that BOINC VIew will run under Wine. No personal experience with that though, my one attempt to get a Linux machine going recently failed ... :( |
|
70)
Message boards :
BOINC Manager :
Download multiple workunits
(Message 2293)
Posted 28 Dec 2005 by Paul D. Buck Post: In the project web site, in your account, in general preferences there is a setting for contact web site every "x" days. THe default is 0.1 days. If you are doing multiple projects as you should ... :) Then this setting should not be very high. *I* have cable service too that can wig out on me and use a setting of 0.5 days. With my project settings I have always had enough work to out last the outages which are usually over in 3-4 hours; and I have nearly a days worth of work on hand (or more) with that setting. |
|
71)
Message boards :
BOINC Manager :
Buying a better Heatsink
(Message 2275)
Posted 27 Dec 2005 by Paul D. Buck Post: Intel's upcomming blitz is planning to do just that. They went back to a PIII based core and are stating over ... So, PROBABLY, the HT will become a thing of the past as it was just a clever way to take advantage of a stalled pipeline. Though, even with a short pipeline if they have the transistors to spare it may be a viable way to milk more performance. |
|
72)
Message boards :
BOINC Manager :
Exiting with Data Left in Memory
(Message 2274)
Posted 27 Dec 2005 by Paul D. Buck Post: Nope, you got it in one. If the application checkpoints on a 1 minute basis you lose, on average, 30 seconds work. Those that have longer, or no checkpointing will cause you to lose all the work to that point. Most project have reasonable checkpoints, but there are gottcha's like the one where I got a "hung" Rosetta@Home work unit that "ate" 25 hours of computer time and got no where... being worked on by the project ... Or the Predictor@Home work units that threw up a FORTRAN error dialog and halted the CPU till Ok was pressed ... as far as I know they are not seriously looking into this problem. The BOINC Client cannot do what the Science Application does not allow it to do. So, the responsibility is back on the projects to ensure "safe" computing. Most do a pretty good job, but, some checkpoints are so large that they are not practical to do very often, like CPDN's, I forget what their interval is ... (15 min?), but, this is one of the reasons I run 24/7 :) |
|
73)
Message boards :
BOINC Manager :
.....
(Message 2273)
Posted 27 Dec 2005 by Paul D. Buck Post: Not going to quote as I wonder it that does not just get in the way sometimes. We do have some guides. Take a look. When you start to realize that some of those are a week or two's worth of work, well, how many guides can you write. Worse, many times it seems people don't want to use some thing like the guides we wrote. "Vision" is good. I am not knocking it. BUt if that is all there is, well, sorry, not much help. My point is that *IF* you stick around BOIC for awhile there seems to be "types" that show up. Sad but true, evolution equiped us with stereotypes to allow us to deal with a complex world. So, start off with the behavior of a "type" and it is easy to get stuck into it. It was not so much politics as different, um, visions of what was needed. Who is to say that mine is right? Or his? This was another point I tried to advance a long time ago that it would be "smart" to set-up some sort of more official community. There is a little movement in that direction. BUt, this is a classic case where perhaps the principles are too close to the problem and without "independent" ways of reconciling the viewpoints we get stuck. I understand Dr. Anderson's position. I just don't agree with it. Well, my opinion of scientists has been shrinking of late. But, it is not likely that my wishes are going to change the world. Doesn't mean I don't try. Lastly, yes, you did come across negatively. Remember, those of us that "patrol" the boards do not take threads in isolation. We all get a reputation. WIth some I am an over enthusiastic cheerleader that cannot see that BOINC is fatally flawed, optimization is always the best option, that the design is fatally compromised because .... In many cases, in *MY* opinion, the problem is not with BOINC, its design, or whatever. It is a misunderstanding of the intent of the design. And in fulfilling that intent, what appears to be a fatal flaw is actually a design choice that does make sense from the system perspective. So, yes, you did come across negatively and that was why the suggestion. Note I did not take umbrage at your comments. But, in the other thread I was explaining the "why" of it. If there was a better way of doing it I would be using it... I just have not found it. Oh, and migration of the data takes months ... I still have not moved all the material off the old site, though usage statictics say that this is not a huge problem. |
|
74)
Message boards :
BOINC Manager :
File.SelectComputer - Connection Failure
(Message 2262)
Posted 27 Dec 2005 by Paul D. Buck Post: @OffBeatMammal, Well, in one thread you are relatively reasonable. Here you are doing what is deplored in the very article you referenced; Flame war. For a flame war to occur there has to be two sides. If I refuse to let you suck me into personal attacks, well, it won't occur. But, here is seems to me that you want to cling to the position that we all hate you and made things such to personally inonvenience you. Trust me, I have better things to do than to think up ways to make life hard on you. But, as much truth there is in that article it ignores the economics. The fact is that open source relies on volunteer contributions means that we cannot control the total quality of the people that volunteer. So, some of us may be a little less than would be tolerated in other venues. Truth be told, BOINC would not have happened if it had not been open source. SETI@Home would have died and then where would many people be? In the two and a half years I have worked on documentation for BOINC I have not yet found a formula for presenting information in a way that suits people's needs. I think I pointed you to my prior site where you can see the source content for the Wiki. Before that was another site, anotheer site, another site ... In each, I was chastized and lambasted for not doing "x" as that was what the agrieved party thought would solve everything. Of course they had no time to help make "x" happen ... As far as the scattering, well, projects are usually funded with many restrictions including what they can spend money on. Well, that may include supporting some of the features you would like. As best as I was able to centralize it, well, the knowledge is centralized. As I said elsewhere, I steal from everywhere, I co-opt as much as I can, and do as much as is possible in my reduced circumstances ... Anyway, this is a flame war and I am not going to get sucked into it ... turn down the heat ... |
|
75)
Message boards :
BOINC Manager :
.....
(Message 2261)
Posted 27 Dec 2005 by Paul D. Buck Post: thanks Paul, Yoiu are welcome. :) Its why I tried ... I understand the work-load and allocation of resource that must go into all this. In an ideal world yes. "Knowing" the other participants as I do, well, I am not quite sure why this transaction turned out badly; and don't think I want to get too deep into that debate. I *DO* know we try to "nurture" as best we can. In my case, well, I have not been able to do much "heavy lifting", nor, to be honest, do I feel that it is encombant on me to do everyones research to provide direct links. Perhaps it is a personal thing. What I do know is that far to many expect detailed, precise, complete, and individual answers. In many cases when that is known to be "impossible". My favorite "bad" example is the question "What is this 'Credit' thing and how does it work?" That question probably has more pages related to it than just about any other page/question in the BOINC universe as it touches validation, cobblestones, benchmarks, quorums, etc. In this specific reference I too missed one point, but I was misled by the way you posed the question. So, my advice was incomplete. I'm sure over time the client / manager evolution will continue - hopefully with input from the real world rather than the ivory tower... sometimes 'pure' function has to be overlaid with a veneer of usability or (even better) user friendliness... Very true. And, if we had time, talent, and additional effort this would be true. But, the sad fact is that with limited resources there is only so much that can be done. The upcoming Account Manager sites may help. There is also a fledgling effort to redesign the interface to something that is easier to use. There is discussion on the developers list if you would like to participate. Not sure where that is going to go, or even if it is going to get off the ground. Sadly, the history of many of the self-proclaimed "experts" is that they are long on complaint and explanation on how if they were in charge that we would be in nirvana. Yet, when you try to pin them down to exactly how things would be changed you get the explanation that they are "vision" level people. Sadly, what is needed is more work and less "vision". I get at least one request a month for edit access to the Wiki. WHen I ask for samples of work, guess how many responses I get. Truth be told, all this stuff is hard. It is just like a beginning discussion on the SETI@Home boards, sadly it is a good idea, sinking under the unpleasent fact that this ground has already been plowed and NO ONE came up with a rule set that couild not be broken, and, in the mean time provided effective relief of the problem ... It would be great to at least make the links to boinc-doc.net more prominent on the boinc.berkeley.edu site (and encouraging the various projects to link to it and contribute knowldege) so people can find the resource without having to dig (and dig and dig!) We have tried. We have even tried to make the Wiki official. We came close. But, I folded when the idea was to: a) start over b) limit the content So, I did not think that this was the way to go and stayed "Unofficial". We have made progress in that almost all projects do give at least passing reference to the WIki as a source. And the volunteers do provide recomendations to people to use it. But, my ability to convince the projects to help me to expand the project specific content has met with limited success. If you want to see what I would like to see from all projects, look at what Chris Randle did from CPDN. I got all the stuff from Classic developed material to standard references they make to threads ... simply wonderful ... And, when I can, I try to co-opt people like Kinhull to take the best of their work and fold it into the Wiki, if possible. We also refer out where possible. I proposed a concept where the BOINC Manager would, on pressing F1 take you to a page directly. Again, if you look at the UCB help page, and what we write in the BOINC Application Owner's Manual you can see why I did not want to limit the content ... :) But that idea too died. I suggested that the Help link be "programmable" so that with a configuration change the participant could point to their help site of choice. That died too. I am not a C++ programmer (my autistic brain can't grasp the notion that the programming language should be designed for maximum obscurity, see the wiki for my thoughts on C++), and not likely to become one. And if I did, well, when I would be able to work I would be adding to the back log of stuff that was not documented. So, I cannot code a change and submit it for inclusion. I *DO* know that if it is good, works and solves issues it WILL be included in the baseline. That being said. e-mail Dr. Anderson and tell him what you think, you may even get a reply. But he is busy so, don't plan on it being a long one. Even better, post on the project or developer's mailing list. Start discussions. But, if you come across as hostile, well, you won't get far. And many good ideas die aborning. |
|
76)
Message boards :
BOINC Manager :
.....
(Message 2247)
Posted 26 Dec 2005 by Paul D. Buck Post: The difficulties in setting up BOINC are known, and work is on-going to try to make it happen. RIghtly you point out that we are not there yet. There is a fairly long list of usability enhancements on the list of things to do. But, these oft times take a back seat to the usability of making it work at all. Many of the recent changes were an attempt in that direction. The fact that we did not quite get there should not overshadow the fact that probably the majority had a relatively trouble-free migration. Cold comfort if you did not have that. I have to admit that I have had a hard time along the way because BOINC has always installed and run for me right out of the box with no issues. Heck, when some people could not run for 20 minutes without a GPF, I was happily runing 24/7 ... The other choices were made on a deliberate, and I think wise basis. Since BOINC is to be multi-project of unknow and ever growing scope it was felt that all but the most vital parts would be "farmed out". In particular, statistics was made as minimal as possible on the basic project sites. For the simple reason that this lets the developers concentrate on the important aspects of the system. This does require that the participant find another place to get any kind of advanced look at their data. And, so far, I have not found one statistics site that does it all for me (yet). So, I use Willy's and BOINC Synergy. Some things I think Willy does better, others, BOINC Synergy. Twixt the two I get what I desire. Because BOINC is open source, it is also impossible for the UCB site to be everything, or even a center point for all projects. They do, again, provide the bare minimum to get us by. But, again, in part economics, in part choice they only make a list of widely known projects that are in production. Documentation *IS* another weak point, for which the volunteer Unofficial BOINC-Wiki is attempting with greater or lesser success to alievieate. In our defense, there are far more hours put into development than documenting, if for no other reason that there are more people that can write code than can write English. There are two places where, perhaps, we may see significant improvement in time for BOINC with regard to single point of access and ease of use. One is WCG which may grow into more than just a project of projects. The other is a Account Manager site that is in development now. Anyway, there was some thinking behind the choices, though it is easy to disagree with the choices made (I too would like to see the ability to make changes to preferences locally). |
|
77)
Message boards :
BOINC Manager :
Managing multiple instances from a single computer
(Message 2246)
Posted 26 Dec 2005 by Paul D. Buck Post: I played EverQuest until my reflexes were too poor that I was endangering my other members of the party. I also got tired of the need to endlessly level the character up to enjoy the content and the only way to do this was to do combat. The feature to tailor an scenario to the group was one of the nicer features, though the manical insistance that grouping was mandatory was the biggest downside for me. Since I never know how long I would have to play it was a bad fit. I tried a couple of the others, and one of them was a little better "fit" for my style, but, I guess I was gamed out ... There are many aspects of BOINC that need significant work, preferences is just one of the small areas. Though, working on the code to make Mac only features will not fly. So, be prepared for the need to be able to make the proposed changes cross-platform. Though with WxWidgets much of that is taken care of for you. |
|
78)
Message boards :
BOINC Manager :
Quad Processor
(Message 2245)
Posted 26 Dec 2005 by Paul D. Buck Post: David, If you are runing more than one project you may want to consider increasing the swap interval to a larger number than 1 hour. If you are running just SETI@Home, don't bother. Welcome to you and your 4 processor system ... |
|
79)
Message boards :
BOINC Manager :
Message says "Not enough disk space", prefs say otherwise
(Message 2233)
Posted 25 Dec 2005 by Paul D. Buck Post: I would respectfully like to suggest that the next major version of BOINC should go through formal usability testing before the user interface is solidified. We need to make it as easy as possible for novice users to join the BOINC community. Yes, it seems obvious doesn't it. In fact, it has been undergoing usability testing for over 2 years. That is why it is so much better than the stand alone DC projects. I can run 7 projects I am interested in all at the same time. The suggestion that the preferences would be better on the client is one that is in the list of things to do. But, I will say that there is some merit in the web based preferences. BUt, over time, new features like this are added to make the system better. But, one side note, much of the work concentrates on the making improvements in BOINC to better support the project's needs. All that said, there are things coming down the pike that *MAY* make things easier ... |
|
80)
Message boards :
BOINC Manager :
Managing multiple instances from a single computer
(Message 2224)
Posted 25 Dec 2005 by Paul D. Buck Post: After trying about 5 different was to present the data I did elect to use MediaWiki as it is the best compromise on features and colaborative editing. Sorry you don't like it. As far as the location of the file, well, the text does say that the file is located in the "BOINC Direcotory", and following that link takes you to a page that says: ---- The directory on your computer that contains all of the BOINC related files. On Windows, this is generally C:\\Program Files\\BOINC On Macintosh OS X, this is generally /Library/Application Support/BOINC Data On Linux, the directory varies, but is frequently ~/BOINC The directory will vary if non-standard choices are made during the installation process. ---- And Bill is correct, if you write something that makes sense and is of reasonable quality I will be more than happy to include it in the WIki. If you have corrections to the text, if they make it better, well, I take all suggestions. I am also sorry you find my effort unsatisfactory, but, like you I too am disabled and for the last couple months my disability has been winning and I have not been able to put in the concentrated effort it takes to do quality work. So, yes, there is a lot of content missing. There is a lot that could be improved. But, there is only one person that was able to put in 8 hour days writing content and that was me, to work on BOINC there are 6 UCB developers, a half a dozen volunteers, and several projects. All of which combine to invalidate the work almost as fast as it can be written. WIth regard to testing of all aspects, well, this is a low budget, open source project. We have to fend for ourselves at times. Welcome to BOINC. |
Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License,
Version 1.2 or any later version published by the Free Software Foundation.