Why let LHC@HOME take over?
Message boards :
Number crunching :
Why let LHC@HOME take over?
Message board moderation
| Author | Message |
|---|---|
Neal Watkins  Send message Joined: 21 May 99 Posts: 192 Credit: 289,148 RAC: 0 |
I'm a happy sharer. 3 projects, each gets 1/3 share. E@H and S@H work well together because the WUs are somewhat the same size as are the due dates. Along comes the last batch of LHC WUs and they all have a due date just a few days in the future. Well...can you say STD and LTD? BOINC stopped working on S@H and E@H until all the new LHC WUs were done. I'm a sharer and I don't like it when other projects manipulate the system for their benefit. That's not what STD and LTD are for. They decided that their project is more important than the others so they set an arbitrary short due date to lock others out and hog the cycles. I guess it's because they are "special" and "important". 
ID: 277216 ·  |
 yank     Send message Joined: 15 Aug 99 Posts: 522 Credit: 22,545,639 RAC: 0 
|
And I thought my machines were messed up. Thanks for the information. http://boinc.mundayweb.com/teamStats.php?userID=14824 ID: 277223 · |
MJKelleher  Send message Joined: 1 Jul 99 Posts: 2048 Credit: 1,575,401 RAC: 0
|
They decided that their project is more important than the others so they set an arbitrary short due date to lock others out and hog the cycles. I guess it's because they are "special" and "important".No, they set their short due date because they've got engineers and scientists waiting in real-time for the results of the work. That's not the case here, where if it takes an extra week for work to come back, nobody's hanging on those results to go forword with their work. S@H and E@H are projects where we're looking for something that (hopefully) has been there for a long time, and will continue to be there, we just want to find it. LHC is designing a new collider, and they need to have a good idea that it'll work before they start construction. This is what LTD is for ... so that one project gets the time it needs to complete its work within deadline, and then won't download more work until that time has been paid back to the other projects that waited. If that's not what you want to see, then LHC isn't a project for you. You got work??? I think I need to visit the Sulking thread, I haven't seen any in weeks..... MJ  ID: 277231 · |
|
Neal Watkins Send message Joined: 21 May 99 Posts: 192 Credit: 289,148 RAC: 0 |
Well...OK...fair enough. BUT The projects exist on the client's goodwill. If something like that is happening, then they need to let their clients know. It is not sufficient to ask me to download their software and be at their beck and call. There is no communication, no agreement, no documentation that this can/will/is happening for this set of WUs and not for others. This 2-way interaction/agreement is totally lacking in this emerging technology. You want to use my computer? OK. What for? How Long? What priorities do you have? What priorities do I have? They didn't ask. They just took. I know it's your important project, but this is my machine and I demand a share in deciding how it's used.
ID: 277238 · |
Astro  Send message Joined: 16 Apr 02 Posts: 8026 Credit: 600,015 RAC: 0 |
I know it's your important project, but this is my machine and I demand a share in deciding how it's used. Neal, What was running when LHC had no work? Einstein and Seti, of course, and a debt was running up for LHC. You set it for a 1/3rd share it wasn't getting, so it was building up a debt that is now being paid back. Also, due to deadlines LHC is running now in priority and a debt for the others is being built up. Your requested "resource share" WILL be followed by boinc. It isn't followed in the short term, but WILL be followed in the long term. This is a BOINC work scheduler issue, not an LHC issue. hope this helps tony  ID: 277263 · |
|
Neal Watkins Send message Joined: 21 May 99 Posts: 192 Credit: 289,148 RAC: 0 |
I know it's a BOINC work scheduler issue. That's why I posted the message here. I think there's a real gray area here. These are real projects. Important projects with deadlines. But, this is my machine and I volunteer it. Don't I get any say in how it's used? My only alternative is to disconnect from LHC. How does that benefit them? My point is that we are not passive users just waiting for anything, any crumb. Perhaps my view is colored by my history in the "business": 32 years as a "techie" and a certified UNIX sysadmin.... (I'm used to being in control...I do things to users, they don't do things to me)
ID: 277279 · |
Fuzzy Hollynoodles  Send message Joined: 3 Apr 99 Posts: 9659 Credit: 251,998 RAC: 0 |
I know it's a BOINC work scheduler issue. That's why I posted the message here. Actually no, the right place to post about BOINC issues is here. I think there's a real gray area here. These are real projects. Important projects with deadlines. But, this is my machine and I volunteer it. Don't I get any say in how it's used? My only alternative is to disconnect from LHC. Yes, it's your computer and your decision how you want to give it's ressources to your projects. So when you decide to participate in a project, where the WU's have short deadlines, yes, they will run and finish before the others. As MJKelleher pointed out, the reasons for the short deadlines are that the results are needed imidiately. E.g. over at Pirates@Home the deadline is very short, so the WU's run with first priority (that they often finish within a minute is another case). And yes, over there we actually sit and wait for any crumb! ;-) So yes, it's our computers, but by signing up to a project, you leave it to the project devs to decide, what work they need to have done, but you can always suspend your WU's, suspend BOINC or exit it. It's your call, if you need your processor power yourself, then you take it, you don't have to ask anybody. And the WU's goes out. Don't worry, they'll run, either on your computer or on somebody else's. And I haven't had any LHC WU's for a very long time! :-( "I'm trying to maintain a shred of dignity in this world." - Me 
ID: 277283 · |
|
1mp0£173 Send message Joined: 3 Apr 99 Posts: 8423 Credit: 356,897 RAC: 0
|
I know it's a BOINC work scheduler issue. That's why I posted the message here. Actually, the scheduler works very hard to make sure your specified resource shares are completely honored. Trouble is, most people look at the scheduler as you did -- in the short term. They say things like "LHC took over" when anything that one project does to get quicker results will result in being throttled later. It works this way: 1) Priority #1 is finishing all work currently downloaded before it expires. 2) Work is downloaded based on actual resources used. If a project gets more than their share of CPU time, then BOINC stops downloading that project until things get caught up. Again, that means that the rules will be bent in the short term to meet deadlines, but it all balances over the long term. ID: 277289 · |
|
Astro Send message Joined: 16 Apr 02 Posts: 8026 Credit: 600,015 RAC: 0 |
I know it's a BOINC work scheduler issue. That's why I posted the message here. Boinc is/was designed for the masses (I.E more than 50% of users). Many don't want to micromanage a client, some do. Overall boinc works fine for "most" users. It's open source if you want to create a client you can control any way you desire. Heck, it might even be something others may want and you could distribute your client. Boinc is trying to get LHC it's third of the pie you requested. It's also fulfilling other functions such as ensuring work gets returned on time. You probably wouldn't have a problem with that, but I've been here since nearly the start of boinc and believe me, many would have trouble returning work on time(as evidenced, by users of Pre 4.35 boinc versions). So, once again, the few are forced to be treated like the many, it's life. You're in a "society" and have little choice but to leave the city, move to alaska and live in the wilderness. LOL You've asked boinc to give each project 1/3rd of your hosts time. I assume you'd like to see One Seti, One Einstein, and one LHC work units downloaded in that order all day long. However, each has different run times, so time is the only way to "balance" the resource share, which may mean, two seti, one einstein, and 3 LHC(depending on the number of turns). If you set your "connect to" setting to .001 days, you're likely to see this blend of work units (as long as LHC has work), The higher your cache the more exagerated the difference of blend. In any case, it will count seconds and balance your requested resource share (when possible, IE LHC out of work) over the long term. tony ID: 277290 · |
The Gas Giant  Send message Joined: 22 Nov 01 Posts: 1904 Credit: 2,646,654 RAC: 0 
|
Set your connect to preference to less than 2.5 days and all should work out pretty well. Most people give thanks to their god when they receive LHC wu's. ID: 277357 · |
Trog Dog  Send message Joined: 18 May 02 Posts: 25 Credit: 208,371 RAC: 0
|
I'm a happy sharer. 3 projects, each gets 1/3 share. Looking at your stats in your sig you crunch/crunched for 4 projects. Looking at your credits LHC has the second lowest score. You have racked up more than 30 times your LHC credits in SETI. If you really begrudge them the cycles detach from the project. The short deadline (at least on my boxes) is a week, hardly just a few days. But really, why post this in the SETI boards, if you have a problem with LHC, post it on their boards.  
ID: 277830 · |
|
Beach Bum Send message Joined: 13 Nov 05 Posts: 178 Credit: 611,717 RAC: 0
|
Perhaps my view is colored by my history in the "business": 32 years as a "techie" and a certified UNIX sysadmin.... Well Neal, I believe you answered your own question about getting to control what is done. Now be a good user and shut up and let the admin do his job.  Beach Bums Current Stats:  Come Join us at Hawaiian Beach Bums ID: 277858 · |
|
Lee Carre Send message Joined: 21 Apr 00 Posts: 1459 Credit: 58,485 RAC: 0 
|
But really, why post this in the SETI boards, if you have a problem with LHC, post it on their boards.the point that quite a few people tried to make was that this isn't a problem, the only potential problem is Neals perception of things, or rather how he thinks they should be done, thus assuming that boinc or LHC are "broken", or as he says "taking over" i strongly suggest anyone with questions about how the work scheduler works has a good read of the BOINC Wiki, starting with the Work Scheduler and as stated, he's free to do what he likes with his own computer, if he doesn't like boinc, he doesn't have to crunch, if he doesn't like LHC he can deplete his LHC cache/queue and detach and besides, there is a standard statement by each project, take a look at LHC's intro page on their site, it answers most of his basic questions, but i suggest that he should have done some reading before starting BOINC if he was in any way unsure of what BOINC is, or does Want to search the BOINC Wiki, BOINCstats, or various BOINC forums from within firefox? Try the BOINC related Firefox Search Engines ID: 277867 · |
The Pirate  Send message Joined: 14 Apr 00 Posts: 191 Credit: 4,929,008 RAC: 0
|
I'm a happy sharer. 3 projects, each gets 1/3 share.  
ID: 279292 · |
MattDavis  Send message Joined: 11 Nov 99 Posts: 919 Credit: 934,161 RAC: 0
|
POW ----- 
ID: 280746 · |
©2026 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.