Message boards :
Number crunching :
Stocking Up On Work Units
Message board moderation
Author | Message |
---|---|
Everette Dobbins Send message Joined: 13 Jan 00 Posts: 291 Credit: 22,594,655 RAC: 0 |
Last outage had me scared so I increased my work units to 6 days worth. Anyone else do this? It really helps the anxiety factor.Hope they can fix this soon. |
Compare Watches Send message Joined: 24 Apr 04 Posts: 81 Credit: 2,862,596 RAC: 0 |
Mines been at 5 days for a while now - so far, not run out during an outage. That message on the front page about Cogent is a bit worrying. Been crunching for Berkeley for a year and not found so much as a single alien - then Cogent come along and casually tell us they've found a UFO. Tactless or what?! Tomcat's Taekwondo Pages |
Pascal, K G Send message Joined: 3 Apr 99 Posts: 2343 Credit: 150,491 RAC: 0 |
One day is more than enough if you diversify. Semper Eadem So long Paul, it has been a hell of a ride. Park your ego's, fire up the computers, Science YES, Credits No. |
Captain Avatar Send message Joined: 17 May 99 Posts: 15133 Credit: 529,088 RAC: 0 |
> One day is more than enough if you diversify. Not everyone wants to run as many projects as we do Pascal Let alone afford an "INTEL DUAL CPU with HT" People will learn to increase there cache to 3 days. I think only once there has been a problem that lasted longer..... Go buy that Ticket Pascal,go to Berkeley to tell em how it should be done.... > |
Toby Send message Joined: 26 Oct 00 Posts: 1005 Credit: 6,366,949 RAC: 0 |
2 days here. Also running several projects. I haven't been without work since the fall of 2004. A member of The Knights Who Say NI! For rankings, history graphs and more, check out: My BOINC stats site |
FalconFly Send message Joined: 5 Oct 99 Posts: 394 Credit: 18,053,892 RAC: 0 |
0.25 days here, so far works fine for me (5 active Projects)... |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
0.1 days with several projects on each computer, and I haven't been dry since August 2004 (only two projects available then and they both went dry at the same time.) BOINC WIKI |
ampoliros Send message Joined: 24 Sep 99 Posts: 152 Credit: 3,542,579 RAC: 5 |
I only run multiple projects on one computer. The rest are for SetBOINC alone because I have a spot in my heart for the project. My only concern about this is that I have one that burns through units very quickly. I recently compiled the whole thing from scratch for this one and didn't think to up the cache in it's group (Home) in case of this sort of thing. Now it's out of work. The rest are just just Windows boxes of varying types and I never even worry about them. I don't like the idea of keeping a big cache because they are always connected but they'll be out of work soon too (2 day cache). BTW - Compiling your own (BOINC and Client) is the way to go. I ran the client under Windows for a while (cached ID) before I converted it to my Linux desktop with compiled BOINC and Seti clients (current ID). I picked up speed over the Windows client. And apparently Whetstone is more important than Dhrystone. 7,049 S@H Classic Credits |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
> I only run multiple projects on one computer. The rest are for SetBOINC alone > because I have a spot in my heart for the project. > > My only concern about this is that I have one that burns through units very > quickly. I recently compiled the whole thing from scratch for this one and > didn't think to up the cache in it's group (Home) in case of this sort of > thing. Now it's out of work. > > The rest are just just Windows boxes of varying types and I never even worry > about them. I don't like the idea of keeping a big cache because they are > always connected but they'll be out of work soon too (2 day cache). > If you run 4.35, you can always set S@H to 10,000 resource share, and some other project (or two) to a resource share of 1. In this case, BOINC will crunch one cache full of the backup project(s) and then never again unless S@H is down. Mind you at the moment 4.35 is Alpha and carries a higher risk. BOINC WIKI |
ampoliros Send message Joined: 24 Sep 99 Posts: 152 Credit: 3,542,579 RAC: 5 |
> If you run 4.35, you can always set S@H to 10,000 resource share, and some > other project (or two) to a resource share of 1. In this case, BOINC will > crunch one cache full of the backup project(s) and then never again unless S@H > is down. > > Mind you at the moment 4.35 is Alpha and carries a higher risk. That's not a bad idea, but you cheat the other project if your other project (of resource share 1) is CPDN or something like it. The CPDN staff ends up sending out a work unit that they may never get back (long cycle time) and the client (us) has no real interest in finishing on time. Or depending on how BOINC is set up (not really sure here), once the reporting deadline for the CPDN project gets close, BOINC will abandon working on the SetiBOINC unit until it completes and downloads anther CPDN unit to ignore. It's a good idea but I don't like "leading another project on." Is there maybe another way? I just noticed that my compiled BOINC client reports itself as 4.68! What does that mean? Maybe I should start paying more attention, I never run the GUI unless there's a problem. I also just noticed that the web site has work assigned to me that I have not recieved yet. Does that mean that my requests are getting through, but they can't send me the work that is assigned? Or does this mean that the connection has been fixed but the network is too swamped to send any work? 7,049 S@H Classic Credits |
Pascal, K G Send message Joined: 3 Apr 99 Posts: 2343 Credit: 150,491 RAC: 0 |
> > One day is more than enough if you diversify. > > Not everyone wants to run as many projects as we do Pascal > Let alone afford an "INTEL DUAL CPU with HT" People will learn > to increase there cache to 3 days. I think only once there has been a problem > that lasted longer..... > > Go buy that Ticket Pascal,go to Berkeley to tell em how it should be done.... > > > > > > > Chill out CA and do I detect DUAL envy.......lololol I can drive to Berkeley from where I live and I am one to never presume to tell the experts anything, but I will voice my opinion as I see fit.....I am only running 2 on two machines and 3 on one, so I guess I should heed my own advise.....lolol The Dual will be in service late next week......... Semper Eadem So long Paul, it has been a hell of a ride. Park your ego's, fire up the computers, Science YES, Credits No. |
Captain Avatar Send message Joined: 17 May 99 Posts: 15133 Credit: 529,088 RAC: 0 |
> > Chill out CA and do I detect DUAL envy.......lololol I can drive to Berkeley > from where I live and I am one to never presume to tell the experts anything, > but I will voice my opinion as I see fit.....I am only running 2 on two > machines and 3 on one, so I guess I should heed my own advise.....lolol > Ha ha... I bet the price of fuel is about the same as a plane ticket out there! We finally went down to 2.11 today... |
John McLeod VII Send message Joined: 15 Jul 99 Posts: 24806 Credit: 790,712 RAC: 0 |
> > If you run 4.35, you can always set S@H to 10,000 resource share, and > some > > other project (or two) to a resource share of 1. In this case, BOINC > will > > crunch one cache full of the backup project(s) and then never again > unless S@H > > is down. > > > > Mind you at the moment 4.35 is Alpha and carries a higher risk. > > That's not a bad idea, but you cheat the other project if your other project > (of resource share 1) is CPDN or something like it. The CPDN staff ends up > sending out a work unit that they may never get back (long cycle time) and the > client (us) has no real interest in finishing on time. > > Or depending on how BOINC is set up (not really sure here), once the reporting > deadline for the CPDN project gets close, BOINC will abandon working on the > SetiBOINC unit until it completes and downloads anther CPDN unit to ignore. > > It's a good idea but I don't like "leading another project on." > > Is there maybe another way? > There is a reason that I specified 4.35. 4.35 will switch into crunch earliest deadline first mode if a deadline is in danger of being missed. Therefore the first WU from the very low priority project will get much more than its share of the resources, but to compensate, 4.35 will not download work from that project for a long time. If you follow through on the completion of the first very low resource share WU, you have not cheated the backup project, and you have set the switch for that project being a backup project (I would NOT suggest CPDN for this role as you will have to wait through the Crunch Earliest Deadline mode for a very long time). BOINC WIKI |
ampoliros Send message Joined: 24 Sep 99 Posts: 152 Credit: 3,542,579 RAC: 5 |
|
jshenry1963 Send message Joined: 17 Nov 04 Posts: 182 Credit: 68,878 RAC: 0 |
Unfortunately, too big of a cache will end up bringing in a lot of work units for many projects, and your crunch earliest deadline first mode will cause lots to go wrong later. e.g. I did the same, set to 10 days, to make sure I had enough seti at home during its down times. But, what ended up happening, since E@H and P@H always have shorter deadlines, even though E@H takes longer to complete than seti, E@H got the priority, and kept crunching E@H, download new E@H, crunch E@H, download new E@H, etc., and the 20 setis I had downloaded, weren't prioritized, until they were the earliest deadline first iems. Meaning the following: I want to crunch more seti, as that is where I am the most interested. I only have E@H and P@H so that I can help out, and also have a backup for when S@H has no more work units, or is down for an extended time. What happens is, by setting cache large, and due to seti's long deadline time, it doesn't run anymore seti for days, almost a week now. Been crunching lots of P@H, even more E@H, and seti hasn't run. It won't run until around the 11th of May now, when S@H becomes the deadline critical. So cache up if you want, I advise against it. My seti stats are paying the price now. Losing many RAC increases due to not running it now, even though priorities are set at 70% seti, 10% E@H, 10% P@H, 10% cpdn. Still hasn't run seti in over 7 days now. I have since cut the cache way back to 2 days now, and just waiting for E@H to get crunched out, and P@H to get crunched out, so that I can get back to my personal favorite S@H. My suggestion: Don't set cache size too high, set it for a few days AT MOST. And let E@H and P@H be the thing your computer does when S@H is really down for a long haul. But who am I, only a user of this system. Thanks, and Keep on crunchin' John Henry KI4JPL Sevierville TN I started with nothing, and I still have some of it left. <img src="http://www.boincstats.com/stats/banner.php?cpid=989478996ebd8eadba8f0809051cdde2"> |
Saenger Send message Joined: 3 Apr 99 Posts: 2452 Credit: 33,281 RAC: 0 |
<blockquote>John McLeod VII, thanks for the idea, I may give that a try. Any thoughts on what project would make a good second?</blockquote> Depends on your interest: WU-size seti for me: ~4h ClimatePrediction: Science goal: predict the climate for the next years WU-size: ~900h (but credit in 72 trickles) Deadline: 1 year Einstein: Science goal: detection of gravitational waves WU-size: ~10h Deadline: 1 week ProteinPredictor: Science goal: Prediction of possible folds of Proteins WU-size: ~1h Deadline: 1 week That's my current open projects. For Pirates (Screensaver development for Einstein), BURP (Big Ugly Rendering Project, pre-alpha status, but open nevertheless) and LHC (Large Hadron Collider simulation, closed ATM) I can't say anything, 'cause I don't crunch them. Gruesse vom Saenger For questions about Boinc look in the BOINC-Wiki |
mikey Send message Joined: 17 Dec 99 Posts: 4215 Credit: 3,474,603 RAC: 0 |
<blockquote><blockquote>John McLeod VII, thanks for the idea, I may give that a try. Any thoughts on what project would make a good second?</blockquote> Depends on your interest: WU-size seti for me: ~4h ClimatePrediction: Science goal: predict the climate for the next years WU-size: ~900h (but credit in 72 trickles) Deadline: 1 year Einstein: Science goal: detection of gravitational waves WU-size: ~10h Deadline: 1 week ProteinPredictor: Science goal: Prediction of possible folds of Proteins WU-size: ~1h Deadline: 1 week That's my current open projects. For Pirates (Screensaver development for Einstein), BURP (Big Ugly Rendering Project, pre-alpha status, but open nevertheless) and LHC (Large Hadron Collider simulation, closed ATM) I can't say anything, 'cause I don't crunch them.</blockquote> You could also look here for more projects: List of all DC projects |
Saenger Send message Joined: 3 Apr 99 Posts: 2452 Credit: 33,281 RAC: 0 |
<blockquote> You could also look here for more projects: List of all DC projects </blockquote> But isn't this a bit more information than asked for? It's about all DC projects, not only Boinc. Do they go together on one puter well? |
Everette Dobbins Send message Joined: 13 Jan 00 Posts: 291 Credit: 22,594,655 RAC: 0 |
<blockquote>One day is more than enough if you diversify.</blockquote> > > I have E@H as backup. I still have 2days worth of seti@home and then E@H will take over. Seti@home is my favorite. I am a little diversified. Captain Avatar is lucky $2.11 a gallon for gas. Where I live gas just dropped to $2.59 a gallon |
Paul D. Buck Send message Joined: 19 Jul 00 Posts: 3898 Credit: 1,158,042 RAC: 0 |
<blockquote><blockquote> You could also look here for more projects: List of all DC projects </blockquote> But isn't this a bit more information than asked for? It's about all DC projects, not only Boinc. Do they go together on one puter well?</blockquote> If you put several projects on one PC they will "fight" over the CPU time. This is a general rule, though some managed to force BOINC and SETI@Home Classic to work on one of their systems. This is one of the big features of BOINC in that a large percentage of the volunteers (Participants) was to be able to donate time to more than one project. Especially critical to people that only have one computer, the ability to do more than one project was minimal. Now, we have a good first selection set of projects with several more in the "open" though in early stage of development. I am also sure that there are even more in work behind the scenes ... |
©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.