101) 留言板 : Number crunching : Highest # of claimed credit (消息 62191) 发表于：8 Jan 2005 作者： UBT - Timbo Post: > > > > > > > the numbers you posted indicate approx 21,000 seconds PROCESSING (not > elapsed) > > time which is around SIX hours > > Subtitled: Making a mockery of claimed credits: OK, just to prove how silly the current credit scheme is for SETI/BOINC, as a short term experiment, I've now changed my SETI/CPDN resource settings to 20%/80% As a result, my lovely 3Ghz P4 now produces claimed credits such as: Result ID: 26064368 Work Unit ID: 80362 Sent: 29 Dec 2004 9:24:18 UTC Time reported: 3 Jan 2005 8:59:53 UTC Server state: Over Outcome: Success Client state: Done CPU time (sec): 55,201.00 claimed credit: 205.38 granted credit: 43.26 Next task - change the resource split to 10%/90% and see if I can claim even higher credits......! And all because BOINC doesn't fully support Win98SE (or is it the other way around ???) Tim

102) 留言板 : Number crunching : Highest # of claimed credit (消息 59364) 发表于：2 Jan 2005 作者： UBT - Timbo Post: > > Upgrade to WinXP Home/Pro and take advantage of the Hyper-Threading technology > built into you fast CPU. I'm sure you CPU time per WU will drop > significantly. Plus, you can do a SETI WU and CPDN WU at the same time with > the HT enabled, no switching between the 2. > > L8R.... > > --- > > My Time: Sunday, 02 January 2005 - 11:04 AM --800 (Pacific Standard Time) > > Hi Siran, Thanks for your kind suggestion. Sadly we live in a world where the US thinks it can rule the world. Software in US might be damn cheap, but in UK, Win XP Pro costs about GBP150 to GBP180 (or about $250 to $300). Hell, some UK dealers are selling it for GBP270 ($500 !!) see here. I know I could get an upgrade version cheaper, but it's still GBP100 to GBP150. Likewise, I'll need to change the mobo for a type that'll support HT and I'll probably need some dual channel RAM and a new HD and........ So, while it's kinda nice of you to put forward your views, I'll pass for now, especially as I'm kinda holding out for a half decent Linux software package that enable me to drop MS totally. Tim

103) 留言板 : Number crunching : Highest # of claimed credit (消息 59357) 发表于：2 Jan 2005 作者： UBT - Timbo Post: > > > > the numbers you posted indicate approx 21,000 seconds PROCESSING (not elapsed) > time which is around SIX hours > Hey "Z". Try this for size - the following is a result I sent in today: Result ID Host Sent Time Server state Outcome Client CPU time (sec) claimed credit granted credit 26064365 80362 29 Dec 2004 9:24:18 UTC 2 Jan 2005 11:30:27 UTC Over Success Done 19,203.03 71.45 pending And this is the processing report via BOINC: SETI@home - 2005-01-02 05:47:49 - Starting result 23ap04ab.24858.25185.504824.243_0 using setiathome version 4.08 SETI@home - 2005-01-02 06:47:49 - Pausing result 23ap04ab.24858.25185.504824.243_0 (left in memory) SETI@home - 2005-01-02 10:47:50 - Resuming result 23ap04ab.24858.25185.504824.243_0 using setiathome version 4.08 SETI@home - 2005-01-02 11:25:15 - Computation for result 23ap04ab.24858.25185.504824.243 finished As you can see: Total "actual" processing time: 60 mins plus 37 mins 25 seconds = 97m 25s = 5,845 seconds Total "claimed" processing time: 11:25:15 less 05:47:49 = 5h 37m 26s = 20,246 seconds Did you notice the disparity? So, not only can't BOINC calculate my actual processing time right, even Win98SE can't pause a process properly either....! And I can't help it if the BOINC credit system is based upon the "claimed" processing time. Tim PS Of course, as BOINC uses some sort of "toss the high and low and average the middle scores" scheme, so if other Win98SE (and other 9x family OS) are crunching the same WU as me, I WIN, because all of our claimed credits (for this WU) would be higher and I'll get MORE than you!!!

104) 留言板 : Number crunching : Highest # of claimed credit (消息 59325) 发表于：2 Jan 2005 作者： UBT - Timbo Post: > Now back on topic: > > Did you notice the title of this thread? > > "Highest # of claimed credit" > > And my opening question was... > > "What is the highest # of claimed credit you have seen?" > > Not show me/us a LIST of your claimed credits then piss and moan about granted > credit! > All I listed was a few WU's showing High # of credit, which is what the original question asked (and that was substantiated by a short list of other WU's). So, I sort of did answer (but in a "fuller" way). What else did you expect? Someone to say "I claimed a credit of 5004.67 on a WU" and to then not back it up. Jeez, you can't please some people.... Have a Happy New Year anyways. Tim

105) 留言板 : Number crunching : Highest # of claimed credit (消息 59324) 发表于：2 Jan 2005 作者： UBT - Timbo Post: > > OT: > > So you yourself say it actually only takes 3 hours per seti WU but you think > you should get the extra credit because the time continues counting when in > another project? Hmmm interesting... Next week I'm going to start taking 4 > hour lunch breaks and continue claiming a full 8 hour work day! I mean I am > still "at work" and the clock is still running. Nope - not once have I actually said "I want more credit coz I'm getting less than claimed." I can live with the fact that some WU's I crunch end up with less than what is claimed. But it would be nice for the project (as a whole) if the credits stacked up better and similar units on similar OS and similar CPU's produced similar credits. Now that would be fairer. Tim

106) 留言板 : Number crunching : Highest # of claimed credit (消息 59323) 发表于：2 Jan 2005 作者： UBT - Timbo Post: > > > > the numbers you posted indicate approx 21,000 seconds PROCESSING (not elapsed) > time which is around SIX hours > Win98SE boxes do NOT show processing time - they show ELAPSED time - this is a known bug in the OS when using BOINC. (Try it yourself on a Win98SE box - even though a WU is "paused" the time continues...!) T. Edit Just found the link to a previous thread on this subject - there's bound to be others !: click here

107) 留言板 : Number crunching : Highest # of claimed credit (消息 59269) 发表于：2 Jan 2005 作者： UBT - Timbo Post: > > > > > > the problem is not that you are getting less than you claim, > > the actual problem is that you are claiming more than twice what you should > for the work performed. > > a normal workunit should get in the 25-35 range > > you are driving a P4 3.06 machine and taking more than 6 hrs per work unit. > > you should be near 3 hrs probably slightly less. > > Do you really believe that you should recieve double the score because you are > doing something to make the processing of a workunit less efficient and taking > more than twice the amount of time you should? > What arrogance and so very helpful....NOT! You are of course making the assumption that I am running SETI/BOINC 100% (eg as a sole project). In my case, on this PC, I am ALSO running CPDN (and until recently) LHC. Hence my WU's DO take longer as they are being "switched" from one project to another every 60 minutes (which is the default). As it happens, each SETI WU completes in under 3 hours of actual crunching time, but due to a 50/50 split with CPDN, so it takes nearer to 6 hours to complete. I can't help it if BOINC cannot work out the calculations properly with regards to the timings on my (or other crunchers) Win98SE boxes....(eg when SETI is 'paused' (while working on CPDN), my SETI project CPU time continues to increase - vice versa with CPDN paused as well!) Timbo

108) 留言板 : Number crunching : Highest # of claimed credit (消息 58999) 发表于：1 Jan 2005 作者： UBT - Timbo Post: > > What is the highest # of claimed credit you have seen? I was poking > around > > today and found this one. > > > > http://setiweb.ssl.berkeley.edu/results.php?hostid=370326 > > Best I got (recently) was: 21591780 5600486 8 Dec 2004 8:21:06 UTC 9 Dec 2004 23:40:17 UTC Over Success Done 25,332.61 92.23 41.33 This was done on a non-overclocked Pentium 4 3.06GHz (non-HT), w/512Mb (although it is running Win98SE, which might be skewing the timing results - think I read somewhere that there was a Win98 "bug" - but my benchmarks are 1602 Whetstones and 4826 Dhrystones). I also regularly get claimed credits in the 60,'s, 70's and 80's, such as: 26064360 6959232 29 Dec 2004 9:24:18 UTC 1 Jan 2005 12:01:04 UTC Over Success Done 22,044.88 82.02 pending EDIT reduced list length to make msg loading quicker /EDIT What really cheeses me off, is that the credit given is so much lower than that claimed.....! Happy New Year to all crunchers. Timbo

109) 留言板 : Number crunching : This is how the perfect WU looks like... (消息 56126) 发表于：20 Dec 2004 作者： UBT - Timbo Post: > I suggested back in the beta that the CPU type, speed, and OS be considered as > a means of classification and got told that this would create a very large and > unmanageable data set. Not having access to the data stored in the host data > table I do not know if that is true or not. Hi Paul, And they mean to say that there current database "isn't" large and unmanageable...given what they are already doing - all I'm suggesting is a simple look-up table that assigns certain WU's dependant on a host specification...! Hells teeth - they'll be telling us next that their database is running on a 'pooter using bubble-core memory...! Timbo

110) 留言板 : Number crunching : This is how the perfect WU looks like... (消息 56125) 发表于：20 Dec 2004 作者： UBT - Timbo Post: > > Maybe the benchmark system should be fixed, so that all hosts claim similar > credit for the same WU. Hi Hans, My suggestion was that all hosts running similar OS and having similar hardware (CPU, Memory) all get *sent* the same WU and hence will return very similar results, which can then be more easily validated (and every-one gets a similar credit). Least it'll stop the current problem of Windows hosts being credited less than the same hardware running *Nix when crunching the "same" WU. Timbo

111) 留言板 : Number crunching : This is how the perfect WU looks like... (消息 55799) 发表于：19 Dec 2004 作者： UBT - Timbo Post: > > It is indeed a beauty as far as timing and validation. Unfortunately > that non > > hyperthreading CPU got screwed on the credit by the 2 hyperthreading ones > :) > > Ah, the joys of the benchmark system! > > > > It happens. There is no perfect system, but I'll give credit where it is due, > this one is pretty good. Eventually, they may seperate processor types (one > of the many someday upgrades). The wishlist is pretty long, and in spite of > the cool factor of credits, I'd rather see the effort go to forwarding the > project (yeah I know this is a part of it). > Hang on a second...! If the SETI host database system knows the type and speed of the host processor (coz it's in our "account info"), what is to stop BOINC only downloading specific WU's to a group of hosts that all have the exact same type and speed of host? That way, everyone who has a particular speed of non-HT Pentium 4 will get the same WU, while those with say a HT Pentium 4, would get WU's froma different batch. Likewise, the type of OS can also be interrogated to ensure that Linux crunched WU's are not compared/credited against Windows crunched WU's. This might then allow the benchmark system to work better. Of course there is a downside - some OS (or hardware) maybe very rare and hence these hosts might not get much credit as there would be few other systems with identical OS/CPU. But as soon as SETI Classic stops, surely there'll be enough hosts with identical spec's to make this only a small issue. Just my 2p's worth. Timbo

112) 疑难解答 : Windows : "To completion" time increasing while "paused" (消息 44924) 发表于：9 Nov 2004 作者： UBT - Timbo Post: > What OS are you running? If it's Win9x or ME, and if you have Leave in Memory > selected in your prefs rather than Remve from Memory, this can happen when > projects swap. May help here too. > Hi Ken, Am running Win98SE, on an ordinary 3.06GHz P4 + 512Mb main memory (not shared) so your assumption was correct. And yes, I do leave the "Leave applications in memory while preempted?" pref set to "yes", but that's only because CPDN suggests that you do this. I don't think any other BOINC project recommends that this setting be set to "yes" but I didn't think it hurts. rgds, Timbo

113) 疑难解答 : Windows : "To completion" time increasing while "paused" (消息 44687) 发表于：9 Nov 2004 作者： UBT - Timbo Post: Hi, I have registered for 3 projects (SETI, LHC and CPDN) and all are running quite well - credits are being issued and work is being done. But, does anyone have an explanation as to why the "To completion" time is increasing even if a particular project is paused while crunching on the same set-up for a different project is being done? Just strikes me as "odd" that the "to completion" times aren't "paused" when that project is not running. Timbo

114) 留言板 : Number crunching : Dumbed down queue parameters (消息 38014) 发表于：18 Oct 2004 作者： UBT - Timbo Post: > I don't want my queue to get as low as four days. On past experience that may > not be enough to ride out a server outage. I want to keep at least six days > of work in the queue. But I don't want twelve days in the queue -- that is > too long. > So, why not "split the difference" and go for 5 days worth (with upto 10 days in the cache) !! Just my 2p worth (English money) Timbo

115) 留言板 : Number crunching : Lost WU's? How are we gonna know? (消息 33738) 发表于：7 Oct 2004 作者： UBT - Timbo Post: > > Anyway, smile, it could be worse ... > Paul: Thanks for taking the time and making the effort to point out all those things to me - things that in fact I did already know. (Not trying to be facetious here, just acknowledging the effort you've made to respond to a message of mine). To other responders: I'm not going to get into any "slanging match" with any user here - it's simply not worth my time or anybody else's - and I have other things to do. I knew what I was letting myself in for when I first joined the SETI project back in 1999. I also crunch data (using BOINC) for other projects, not just SETI (see my signature - if it's working !! - I'm not "SETI-centric"). Now that other projects are coming online, it'll be interesting to see how they scale up, and face being used by tens, hundreds or thousands of users. I am sure that unless these projects maintain a professional and responsible approach, then they will undoubtedly lose a large number of volunteers simply because they could not cope with the users demands. And that's sad, because fundamentally, I'm sure that the majority of BOINC supporters are in it for the data and not the score sheets. But until the people running these projects realise that they can *only* rely on volunteer resources *if* the project is run responsibly, then I'm afraid that some of these projects may well fail due to poor management and poor infrastructure. And that will be a far sadder loss, than any "apology" offered on a project front-page. As for me, I "crunch" coz I like to contribute something to a worthy cause - no other reason. I "moan" coz sometimes distress can be caused due to an avoidable failure - and that needs to be dealt with. And I write this because I can - not to offend, but to ensure prompt action can be taken. And if puerile comments get made as a result, then tough - sh*t happens I know. But some-one has to clean it up. Timbo

116) 留言板 : Number crunching : Lost WU's? How are we gonna know? (消息 33590) 发表于：7 Oct 2004 作者： UBT - Timbo Post: > > As a "SETI "cruncher" who is supplying free CPU power and are paying for > > electricity to run my PC's to help this project > > It's crap like this that I was ranting about in that other thread. > Papa, All I was doing was pointing out that the SETI project has a "responsibility" to the people who donate their equipment time and resources. So far, the new project, despite the "growing pains" has NOT been that responsible in terms of ensuring that donated work is correctly processed. If they had been responsible, they should have waited until the project was more stable. That's not to say that they wouldn't have still had a few problems, but at least they would be more accountable. After all, you wouldn't trust an online bank if their back-end systems didn't take care of your money. All I'm stating is that the responsibilty for taking care of "hard worked for donated resources should be of 'similar' importance (to that of an online bank). And the odd shrug of the shoulders and a bland "sorry" isn't really good enough ! As for this being "crap" - long live democracy and free speech !! Timbo

117) 留言板 : Number crunching : Lost WU's? How are we gonna know? (消息 33422) 发表于：6 Oct 2004 作者： UBT - Timbo Post: Attn: Webmaster / SETI head honcho I just saw this posted on the SETI/BONC home page: "October 6, 2004 There's been a problem with file uploading for the last few days. We think it is fixed now. Our apologies if you lost work because of this." As a "SETI "cruncher" who is supplying free CPU power and are paying for electricity to run my PC's to help this project, I'd just like to know how on earth you expect us to know that we've "lost work" when YOUR system has failed (especially as you've turned off the "view results" facility). Give us a break will you? - turn on the "view results" and give us the ability to log ALL work done, so WE can compare the results we achieve on our own PC's with the results you think we've sent. Then we'll know what good we're doing and whether we're wasting our time due to your hardware/software failures. Timbo <img><img><img>

118) 留言板 : Number crunching : Ready to send 1,657,042 - BUT MESSAGE FROM SERVER- NO WORK! (消息 32137) 发表于：2 Oct 2004 作者： UBT - Timbo Post: > > The more projects you attach to, the fewer WUWUs you get for each project, > think about it!!!! > > > My name is Pascal and this message has no meaning, but still has my > approval... > Pascal, You are of course correct that if you crunch WU's on one project for 100% of the time, of course you'll get more WU's than if you only crunch say 50/50 on two projects. But I'm getting say 1 SETI WU a day (and no cached WU's), my prefs are set to crunch 40% for SETI and I'm running a P4 PC at 3GHz, which should return maybe 7 a day at 100% (or just under 3 a day at 40%). So, why can't *I* ensure that my cache has enough WU in order that the PC is kept working if I have no other WU to crunch and if SETI goes offline??? Timbo

119) 留言板 : Number crunching : Ready to send 1,657,042 - BUT MESSAGE FROM SERVER- NO WORK! (消息 32131) 发表于：2 Oct 2004 作者： UBT - Timbo Post: > Hi Timbo, > > I would let the time out expire, and if it doesn't automatically update, > manually update it. > > I run the boinc_gui with the following switch (-return_results_immediately). > This maintains a full cache at the desired setting. In theory it downloads a > wu for everyone you send back. In practice it sets a low and high water mark > with your desired cache as the median value, ie low=10days desired=11days and > high=12days. When your cache reaches the low mark it usually tries to fill to > the high mark. > > In theory ;-) > > Dunc To all who've responded. Thanks for your messages and help. Seems that I'm not destined (currently) to be able to re-fill my cache to the level I think I need, in order to prevent a lack of WU's from affecting my ability to crunch. (I've tried to prevent that anyways, by joining a few projects, but given I've been crunching for SETI since August 1999, SETI@home has always been my "preferred" choice for using spare CPU cycles). So, I'm going to try and ride this one out - and hopefully I'll get enough WU's to keep my credits rolling (as you can see from my signatures, I've been quite succcessful on SETI/BOINC up to now - only had "trouble" in the last few days) and I cannot explain why I'm not getting WU's in any significant quantity. C'est la vie ! Timbo PS Just did a manual update and SETI says: SETI@home 2004-10-02 18:53:14 Message from server: No work available You just can't win !!

120) 留言板 : Number crunching : Ready to send 1,657,042 - BUT MESSAGE FROM SERVER- NO WORK! (消息 32096) 发表于：2 Oct 2004 作者： UBT - Timbo Post: > Hi Timbo, > > It may be worth increasing it to 2 for example and seeing what it does. You > can always put it back afterwards. > > I personally run mine at 11 days! In the past it has served me well during > outages. > > Dunc Dunc, I just changed it to 5 and left it for a while (before I answered your last post). Now when I do a manual update I get: SETI@home - 2004-10-02 16:27:29 - Message from server: Not sending work - last RPC too recent: 546 sec What's an RPC? And how can it dictate that "it" doesn't want to send me work? I thought that the BOINC client allowed for greater "user preferences" compared to SETI Classic !! (BTW: I used SETI Classic and SETI Driver, so when things got "tight" on the SETI server, I increased my cache and it worked flawlesly. BOINC now seems to take more control *away* from the user. Timbo

前 20 · 后面 20

 
©2020 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.