101) 留言板 : Number crunching : why oh why (消息 55723) 发表于：19 Dec 2004 作者： ric Post: @mikey people outside US and running dial in, they dial to their local internet provider with local costs. Nearly nothing by night.. It's not a BBS, it's internet connection, even on analog dial in. --> every effort to get a faster validations is welcomed.

102) 留言板 : Number crunching : why oh why (消息 55717) 发表于：19 Dec 2004 作者： ric Post: There is no problem to make more than 20 30 or 40 Seti Work Unit a day. Expand your thinking: what are *some* people doing with the CPUs? They are over clocking a much as they can. A to much OC CPU IS doing garbage and a Work is "finished" in less than let's say 10 Minutes or much less. So as humans are, some are just doing it, letting the work go broken, other are more smart and they disable internet connections and are running SAVES to avoid data loss. That was just one example how *everybody" CAN doit. Sure with optimized code, a lot of gain CAN be possible. We all can see the relatively deap cpu heat while crunching seti WUs. Other Work from other project, is/was demanding the same CPU much higher And the tomorrow's CPUs will be much faster than todays (dual core) 16-18 Seti Work Units is not to much estimated for a upper 3.0 Ghz Presott the 765 socket 3.6 GHz or 3.8 GHz Intel chips will do with *normal* clients over 20 Seti WUs a day

103) 留言板 : Number crunching : Boinc for multiple CPUs. When will it be fixed? (消息 54419) 发表于：16 Dec 2004 作者： ric Post: >Boinc for multiple CPUs. When will it be fixed? it more looks like the fix(es) must be done on your client side. *my* HT based clients are doing the job very well. With several client releases. (os w2k / XP home + server / Windows Server 2003) Excluding the delayed validation at seti, I cant find a reason to complain

104) 留言板 : Number crunching : Only one CPU being detected? (消息 54247) 发表于：15 Dec 2004 作者： ric Post: > I just tried un-instaling BOINC and then re-installing it, and running the > benchmarks again, and it still says "Number of CPUs: 1". *sigh* When you open the taskmanager, is it showing 1 physical and 2 logical CPUs? If not, verify in the PC setup (Bootphase-->setup) if the HT mode is "enabled". goodluck!

105) 留言板 : Number crunching : RAM effect (消息 53654) 发表于：13 Dec 2004 作者： ric Post: What would be the change in speed when talking about finishing 1 WU, if you had lets say 512MB of ram, instead of 256MB ? In the past, on a AMD 2400 have had 256 MB RAM DDR, ready for for 266 MHz. The same speed as the cpu is running (fsb multiplicator) The memory was replaced later with a single stick of 512 MB, ready for 333 MHz The MSI Mainboard was able to run the memory at the higher speed. In addition, the memory timing of the 333 MHz stick could run stable with 2.0 instead 2.5 It was seti classic, the average crunching time from about 3:30-3:45 could be reduced to 3:08, depending the AR of WU As the previous poster have said, not the amount of the memory is in the foreground, but much more the finaly resulting (stable) operation speed. If you have a Mainboard, supporting dual channel memory, it could be a gain, due the memory bandwidth is added with the 2nd stick. For instance, running a Intel based HT ready mainboard, with only one memory stick on board (one channel used) can decrease the resulting performance in a drasticaly way. Also had on SETI classic, on a 3.0 GHZ HT, 2 sticks (two memory channels) gived 14-16 WUs per day. For testing, I took the 2nd Memory stick away (one channel modus), and the same system with the same client was reduced to 10-12 WUs a day. I dont know, on the AMD side, with their HT (hyper transport) a 2nd stick is mandatory to have the high level memory bandwidth. Perhaps one last thing, just to reflect: Even a "extrem fast" system (with expensive ram and highend components) IS slower than 2 or 3 medium or low cost CPUs. Taking all low cost of the system together, the totaly bill will be deeper too. At ebay, have seen "old" P3 PCs, ready for use, starting at a price level less than 20 Us Dollar. Including a m$ license... An new m$ license is more expensive as the full system. Or have seen some AMD 2400 based PCs, starting at about 30 US dollar. take 2 or 3 of them and give the lady prescott a "bad time" ;-) Still trying to understand how boinc is validating the work done: as more I'm looking deeper into the stats/results, just in point of view "getting credits", as more I' m convinced that 2 AMDs (or intel) running at about 1400 1600/1800 will return much more "credit" than one single 3.xx GHz engine. It's not a technical question, it a question how finaly the resulting credits are taken and calculated. To fast CPUs will most of the time get LESS than estimated credits, the other way, to SLOW CPUs will also get LESS than estimated. Those are my personal obsevations/reflections. If you start to compare you will might find the same behaviour.... so finaly if you want to spend money for "more crunching power", spend it in a "smart" way. Could it be, your laptop is NOT running at the full CPU speed, or the CPU speed is trotteled down? friendly ric

106) 留言板 : Cafe SETI : CANADA HAS HOMO MILK! HOMOMILKPHOBES STAY OUT! (消息 53292) 发表于：12 Dec 2004 作者： ric Post: A while ago, human have been called homo erectus look here for homo erectus

107) 留言板 : Number crunching : Thread closed! (消息 53181) 发表于：12 Dec 2004 作者： ric Post: Hi Voyager > I want to know when I'll hit 100k ... 99% in about 12-14 Days, taken your RAC. Early January > I liked Stan Popes page that tells you when you'll hit a certain milestone. > Will it be long befor we have something like that here? Probaly yes. Missing Stan Pope's pages too. Even when it can figured out what's the daily throughput on the client side is, the validation is still the dark side in the formula. To much estimates to take. If your crunching "power" doesn't change, You can take the received number of credits in one week and you have a weekly estimate. Not all weeks passed, the SETI project was able to run a full week without having any saturations. But as more work is done, as less is left to do, like at seti classic. When you get 99,99934 % done, what else can be done? You will never be able to do more than 100%... I'm very glad, seeing other (boinc) based projects emerging out. friendly ric

108) 留言板 : Number crunching : Thread closed! (消息 52858) 发表于：11 Dec 2004 作者： ric Post: > Yeah, I went over 10K in credit today! Wondering when I will go over > 100K.... > Congratulation! good job, keep on!

109) 留言板 : Number crunching : CPU speedlimit :-) (消息 52604) 发表于：10 Dec 2004 作者： ric Post: some of Todays Work, final cpu time all clients gui 4.13, gui minimized AMD 2800 regular at 200 MHz,256 MB RAM w2k last seti times 2:29:03 2:29:08 2:32:02 AMD 2500 at 2190 MHz FSB 197,256 MB RAM w2k last seti times 2:36:14 2:36:02 2:35:36 AMD 2600 at 2240 MHz, FSB 173, 256 MB RAM w2k last seti times 2:51:14 2:51:22 2:51:22 Intel Prescott 2.8 at 2.940 512 MB HT on w2k last seti times 3:23:37 3:25:19 3:25:44 (takes 2 Wu at time)

110) 留言板 : Cafe SETI : testing vivitar only (消息 52566) 发表于：10 Dec 2004 作者： ric Post: > Putted it away, due turned to FD mounted devices > FD? > Yes FD. First was M42, then FD and now we use EF. Perhaps missed one;-) Emm, to avoid possible confusion, talking about (D)SLR Canon optical devices, mounted on a EOS based body

111) 留言板 : Cafe SETI : testing vivitar only (消息 52438) 发表于：9 Dec 2004 作者： ric Post: vivitar?? Had a 80-200mm Zoom, Canon FD mount from Vivitar. Long time ago, with manual focus only. Putted it away, due turned to FD mounted devices @cc: don't talk/post about, DOIT!

112) 留言板 : Number crunching : Buffer Overrun Error...!!! (消息 52391) 发表于：9 Dec 2004 作者： ric Post: > I was getting this error on P-4 3.2GHz Prescott overclocked fsb/225Mhz I > backed it off to fsb/223Mhz and quit getting error. > ========= 223/225 fsb, please correct, it looks like the cpu is running about at 35xy MHz. Depending the Chip "quality" the speed of 3.2 is garantied Take a look to the memory timing, SPD or not, and the values. There is still a quality range ob level sticks, some are able to run fine and stable with ABOVE the recommanded speed settings. When I take a look, with tools (Asus) the max speed for 3.2 Prescotty is 3733 MHz.Without taking precaution, higher speed can/will damage the cpu. This doesn't means, the PC will run stable with this speed. And with every apps. But PoorBoy, Your values ARE below the theoretical max. This puttes the focus, a setting (Bios Level) and/or a Component, a part is "not running well". It's not sure, it must be the CPU itself. After verifying the BIOS Values, You can exchange the MEMORY stricks, to exclude a possible RAM "low tolerance" issue. An other problem center could be, the powersupply, not able to give STABLE and continious/on going "power". I know U are running watercooled based system. Will in this case, the chipset be cooled enough On the other side, even decreasing the fsb to 210/215, finaly it will only take some minutes over 2:30 or so, so the "gain" with higher speed is "relative" Do you remeber the good old day at LHC? There where we discovered, to much OC systems returning bad or 0 validated results? It looks like, here at seti this issue is not so important. BTW what a your time with seti? *MY* 3.2 Prescotties are about 2.20-+ (with kingston low speed DDR) and 2:45-+ with regular apacer DDR both are running at speed 3519, fsb 220 AND stable. For LHC they have to downclocked to 3380 or even less. Greetings ric

113) 留言板 : Number crunching : LHC Update (消息 52379) 发表于：9 Dec 2004 作者： ric Post: Thanks for the Info Matt GR ric

114) 留言板 : Number crunching : Out of WUs (消息 51985) 发表于：7 Dec 2004 作者： ric Post: > Errrrrrrr, Sorry ... ;) ================================== Schmock!!! this means u where dried out? no more work to do? To fast CPUs U have! friendly ric

115) 留言板 : Number crunching : Time Zone Questions (消息 51833) 发表于：7 Dec 2004 作者： ric Post: >I wear a watch and have no wall clock Yes, but only the big wall clock makes the sound of coucou (KuckKuck)

116) 留言板 : Number crunching : What are the pros and cons of joining a team ? (消息 51727) 发表于：6 Dec 2004 作者： ric Post: @oatfire-san konitschiwa! Hajimemashite, Nihongo ga wakarimasen. Wakarimasu ka? Nihongo de nanto iimasu ka: WELCOME ABOARD! goodluck with SETI and enjoy the trip Yoi ichinichi wo, mata aimashou! ric

117) 留言板 : Number crunching : Time Zone Questions (消息 51616) 发表于：5 Dec 2004 作者： ric Post: <quote> Alpha Clock is an tiny clock utility with miniature, but very nice LCD-style watch display. You can choose any colors for display background and digits, or use any predefined color scheme. On Windows 2000 or Windows XP also possibly to set the transparency for watch display. Alpha Clock can display your local time or UTC (Coordinated Universal Time). </quote> <br /> Alpha Clock get it from here <br />

118) 留言板 : SETI@home Science : New World Network Speed Record: 101Gbps ! (消息 51468) 发表于：5 Dec 2004 作者： ric Post: take a look here, a work view of *some* cross ocean Internet trunks can be shown, closer go into by clicking Global Network those a "public" networks, I wouldn't be surprised to see some more, not for the grands public, but military used only trunks. Expensive, but also working for continent-continent connection, is the way over satelit.

119) 留言板 : Number crunching : herr david anderson ihre zeit ist abgelaufen ! (消息 51464) 发表于：5 Dec 2004 作者： ric Post: Arnold! In your profil your are talking in french. Your german doesn't looks like done by nativ spoken German, IMO Alhors, excuser-mois pour ma grammer miserable, mais il me semble pas nessesaire, d'offrire un francais parfait pour passer le message. Nous tous avons des moment faibles et la confience dans le project de BOINC,(est come il est gerer) c'éloingne. Mais c'est dans nôtre esprit que cela ce passe. Tu as surement vus, il font de leurs mieux, en tout cas, ils essayent. (ils=berkeley) Cela ne dit pas, qu'ils font leurs mieux dans tes eyes. Tu sais ce son des people d'une universitaire, cel an'est pas la vis réelle. Il y a exactement 2 chemain à prendre: - rester et axepter la façon comme cela ce pass aujourd'hui - ou quitter. Riens d'autres:-( Comme on n'as pas grands le chois, il faut si mettre et garder une distance et pas trop au sévère. Il y a bien plus de problème plus graves dans la vi. Heureusement boinc n'est pas limiter a "Seti only", LHC, si ils ont du travails et les pirates peuvent donner aussi beaucoup de plésire et du fun. C'est à tois de gerer la puissence de tes ordinateur... Fais une pause et reviens voire si tu n'as pas perdu le gut de SETI A bientôt ric

120) 留言板 : Number crunching : Scheduler is back online (消息 51452) 发表于：5 Dec 2004 作者： ric Post: > It seems that LHC has hijacked the seti-project: This client was *hijacked* by cpdn: > >--- - 2004-12-05 00:56:06 - Starting BOINC client version 4.53 for windows_intelx86 Pirates@Home - 2004-12-05 00:56:06 - Host location: home Pirates@Home - 2004-12-05 00:56:06 - Using your default project prefs SETI@home - 2004-12-05 00:56:06 - Host location: home SETI@home - 2004-12-05 00:56:06 - Using separate project prefs for home Pirates@Home - 2004-12-05 00:56:06 - Host ID is 4293 SETI@home - 2004-12-05 00:56:06 - Host ID is 368931 --- - 2004-12-05 00:56:06 - General prefs: from unknown project http://climateprediction.net/ (last modified 2004-11-27 16:23:50) --- - 2004-12-05 00:56:06 - General prefs: using your defaults SETI@home - 2004-12-05 00:56:06 - Resuming computation for result 09ap04ab.28471.13618.1028410.116_1 using setiathome version 4.08 --- - 2004-12-05 00:56:07 - Insufficient work; requesting more Pirates@Home - 2004-12-05 00:56:07 - Requesting 561600.00 seconds of work Pirates@Home - 2004-12-05 00:56:07 - Sending request to scheduler: http://pirates.vassar.edu/cgi-bin/scheduler Pirates@Home - 2004-12-05 00:56:11 - Scheduler RPC to http://pirates.vassar.edu/cgi-bin/scheduler succeeded this client was never attached to cpdn! 2 other m$ Clients are attached indeed to cpdn but not this host.

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.