|
1)
Message boards :
Number crunching :
Jukka Klem, please fix the validator / database problem?
(Message 12518)
Posted 27 Jan 2006 by itenginerd Post: I could give at rats butt about the science I'm only in it for the credits haw haw. No credit no participation lol. At least you're honest. Hey, will you do work for me if I promise to provide you an ample supply of imaginary numbers? I could really use a good lacky these days... Especially if it doesn't deplete my hoard of imaginary numbers I've got here to dole out. (j) James |
|
2)
Message boards :
Number crunching :
going, going.........
(Message 12356)
Posted 24 Jan 2006 by itenginerd Post:
lol Paul. As long as you can recognize and fess up to it, that's all I ask for.... deeeep breath. 8) (j) James |
|
3)
Message boards :
Number crunching :
New work when?
(Message 11958)
Posted 14 Jan 2006 by itenginerd Post: Why not? I knew as little about this as anyone else. ;-) lol. That's what it was, Chrulle. They had to wait until the project admin least expected it, then drag you in from your otherwise peaceful weekend to work. Nice to see it's not just my users that do that to me. 8) (j) James |
|
4)
Message boards :
Number crunching :
Work available, but project is still down!
(Message 11956)
Posted 14 Jan 2006 by itenginerd Post: How about inviting those guys into this board, to show them what turmoil is caused by their lack of communication-culture? ;-) 1) The accelerator engineers have no investment in this forum as far as PR goes. I sincerely doubt they care a great deal that you're as ramped up as a three year old on birthday cake about the work. 2) You're as ramped up as a three year old on birthday cake about the work available. We've been stoic and patient this far people. Let's have a deep breath here, eh? (j) James |
|
5)
Message boards :
Number crunching :
Pending credit never approved
(Message 11829)
Posted 4 Jan 2006 by itenginerd Post: Don't I remember hearing rumblings about LHC's over-eager deleter that did this exact same thing last year? It should have been fixed by the time of this WU, but it's identical behavior. Once the quorum was reached, the deleter was removing the quorum results. So the result in question would be an 'antique'--the exact kind of result that caused S@H some of their problems last summer... (j) James |
|
6)
Message boards :
Number crunching :
New work when?
(Message 11823)
Posted 4 Jan 2006 by itenginerd Post: Cool ... someone took the time to count them ... :) Nah--I don't have the time to count 'em (and not enough fingers!). I took the ones I could count on one screen, squared it twice, and multiplied by my RAC. So I only underestimate the real number by a million or so. 8) Sometimes I wonder what is so hard to understand about the concept that "projects will not always have work"? Nothing. I find that usually, people just get mysteriously antsy when there's no results coming from a project they like. You know the feeling. You all feel it right now--that's why you're reading this. Something's wrong, and you can't put your finger on it. You just know it goes away when your boxen are crunching again. Some people just lack the brain-to-fingers filtering capability that keeps people like you from writing 'Where are my WU?' rants. Either that, or you're worried that someone else is getting ahead of you in the Great Imaginary Number Race. Your pick.... (j) |
|
7)
Message boards :
Number crunching :
can't download
(Message 11794)
Posted 3 Jan 2006 by itenginerd Post: 1) If you want to shortcut the LTD figures, just detach, reboot and reattach. That should clear the LTD for the project. Obviously not something you want to do weekly, but is good enough for an extended outage like this. 2) Why do we always refer to debts as averaging zero? Wouldn't it be as correct (and more precise) to say they sum to zero? The two statements are identical, but I find myself preferring the latter. Is there a specific reason we do one or the other, or was it just preference and wiki-inertia? 8) I nitpick--I got nothin' else to do while I wait for my WUs. 8) (j) James |
|
8)
Message boards :
Number crunching :
New work when?
(Message 11793)
Posted 3 Jan 2006 by itenginerd Post: I hope not because I only signed up last month and I haven't had any work yet. :-( lol. Timing is everything, eh? Hey Paul, I think you're overreacting--I only counted 275,326 idendtical 'when will I get my work' threads on the S@H boards--not including attempts at normal threads that just got hijacked... 8) I'm watching the front page, too. Ready to get LHC unsuspended so my LTDs can get back under 42 million... (j) James |
|
9)
Message boards :
Number crunching :
project priority ideas.
(Message 11442)
Posted 22 Nov 2005 by itenginerd Post: 1) John means that he's the one writing the panic mode codeNot only the panic-mode; as far as I know he has developed the complete actual version of the local scheduler: Correct. I didn't mean that EDF was John's only contribution. I just slid past the rest for brevity's sake... 8) (j) James |
|
10)
Message boards :
Number crunching :
project priority ideas.
(Message 11438)
Posted 22 Nov 2005 by itenginerd Post: Don't know what you mean by: 1) John means that he's the one writing the panic mode code 2) Someone correct me if I'm wrong, but I think the problem with a 24-hour value for 'switch every' would help DEFEAT the panic mode code. If what I remember is right, the EDF scheduler only kicks in when results are switched, i.e. at the end of the 'switch every' timer. So by going to a 24-hour value, the panic mode could only possible kick in A) every 24 hours or B) every time a work unit is done (or C) every time you hit Update manually). If you have a slow system crunching that takes more than 24 hours to crunch a result, then you would very likely miss deadlines. Granted, this is founded on the ideas that: 1) you have a slower machine or large results, and 2) that I'm right about EDF and how it kicks in. Can anyone affirm or correct this for me? (j) James |
|
11)
Message boards :
Number crunching :
I am stil missing credit for 1 wu.
(Message 11434)
Posted 22 Nov 2005 by itenginerd Post: I don't think you will get any credit for this workunit. You have finished it in half the time of the other processors while they are a lot faster than yours. This suggests that your client has encountered an error, this can happen of various reasons and if it doesn't happen all the time it isn't a big problem. On one hand, I agree with alpina--it's odd that your time is half the next lowest time. On the other hand, that result should have validated one way or the other. If there was a problem, it should have showed up as Validation State: Invalid instead of Initial. I can't think up a good reason why you didn't get credit for this result. (j) James |
|
12)
Message boards :
Number crunching :
Host corruption solved?
(Message 11064)
Posted 27 Oct 2005 by itenginerd Post: @Chrulle - Had some problems this morning (~12:00 UTC), but I think that was left over from overnight/yesterday. All appears well now. No host corruption after 8 uploaded results just now. Great job--looks like it's working fine! (j) James |
|
13)
Message boards :
Number crunching :
Host corruption
(Message 10943)
Posted 25 Oct 2005 by itenginerd Post: On my initial look at my hosts, nothing was awry. I manually updated to report and pull work, and that host went corrupt (I had apparently used my quota of 200 WU/day--I don't get close to that!). Another manual update brought things back to normal, and I was able to download work again. A couple minutes later, I suspended Seti@Home, and LHC downloaded another few WUs. No corruption there. That sounds to me like the problem is only in the report uploading portion of the cycle. Or else it has to do with the manual update. I've set NNW for now to see if I can't force BOINC to do nothing but upload results. We'll see what happens. I run CC 4.72. (j) James |
|
14)
Message boards :
Number crunching :
Can anyone enlighten me?
(Message 10788)
Posted 19 Oct 2005 by itenginerd Post: <blockquote>Over success done only means that the Science Application did not abend. Because the SA "thought" the work was done correctly, it is reported as "Success".</blockquote> holy hell, Paul. What are you--an old Netware admin? --twitch-- Here I thought all was right in the world for a little while, and you had to go and say that. I hope you're proud of yourself. 8) ON POINT: it looks like whatever the offending host's problem(s) were, they've cleared up, because it's returning valid results again. He only went through 270 results in two days. Yikes. (j) James |
|
15)
Message boards :
Number crunching :
Workunits Coming Soon?
(Message 10752)
Posted 16 Oct 2005 by itenginerd Post: <blockquote>Yeah sorry about that but i am on vacation in Denmark, so i have not had time to follow up on the status of work.</blockquote> -frustrated sigh- The only thing worse than a systems admin on vacation is a systems admin <strong>on vacation in Denmark</strong>. Seriously, Chrulle. You CERN academic types need to get with the program here and show some dedication to this project. Here in the US, we work. All day. Every day. Even on our vacation. No exceptions. Step up and get the job done, man! (for those who haven't picked up on it yet, the above is a joke. Take it as such.) In reality, I'm jealous. Wish I could get a vacation. Maybe this moving to Europe idea isn't so bad.... 80% taxes, but 6 weeks off a year.... Hope you had/are having a good holiday, Chrulle! 8) (j) James |
|
16)
Message boards :
Number crunching :
New kind of workunits?
(Message 10647)
Posted 8 Oct 2005 by itenginerd Post: <blockquote> James, I am not sure I understand your question. </blockquote> I was kind of pondering a justification framework by which I could request a BOINC-style 'get sweet WU' option'... It was more thinking out loud than anything else.... (j) James |
|
17)
Message boards :
Number crunching :
New kind of workunits?
(Message 10642)
Posted 8 Oct 2005 by itenginerd Post: <blockquote>That's right, but all my wjun4C_v6s4hhpac WUs were finished between 6 and 12400 seconds, the w5_lhc_coll_IP15_trip_meas WUs are taking up to 22300 seconds. </blockquote> You're right. They take me about two and a half times longer than a normal WU. I got a set today that wrapped up in under 30 seconds. 16_18 amplitudes never last long... 8) Clearing those two short units earned me 5 more WUs--thank goodness that correction factor they put in 4.72 cuts both ways. I've been getting LHC WUs that say they'll take 8 hours for a regular WU, which is double my normal times. @Paul - Interesting philosophical question: Wouldn't it just be a preference issue as to whether a user wants a short WU (less BOINC credit) or a long WU (more BOINC credit)? That would give me less overall credit, since I'm chewing up 100 0.1 credit WUs instead of one 20 credit WU. Gives the number-hugry a chance at more credit, and me a chance to fill my dream of actually using all that DL quota I've got laying around.... 8) (j) James |
|
18)
Message boards :
Number crunching :
Claimed credit changed?
(Message 10636)
Posted 8 Oct 2005 by itenginerd Post: Shouldn't the result, its claimed credit, and its granted credit still show up in the list of results? I realize that the results page gets shortened when a WU/result is assimilated and deleted, but I've got about 80 results in my list dating back to Aug 31, all of which are finished with validation. Can't see any reason why they wouldn't be assimilated and deleted... (j) James |
|
19)
Message boards :
Number crunching :
New kind of workunits?
(Message 10567)
Posted 3 Oct 2005 by itenginerd Post: <blockquote>One of the problems with LHC@Home is that the work unit can end at almost any time. I once tried to see if I could make a histogram of the results so I could get better average times on the turn counts ... no such luck... Now that I have a better "feel" for the work unit contents I may try again ... the only thing that may be a problem is that I am not sure that the "order number" is in the same character position ...</blockquote> Good old engineering--no need to spend any more cycles if 'x' happens. The only way I can track which WUs may hop out early is by the amplitude numbers. They're always even; I see 4_6, 6_8, and 16_18 in my queue right now. The lower the amplitude number, the higher the chance to go full term. Also, once you get up around 8_10, the WU processing time will get shorter as well. I'll be surprised if that 16_18 WU goes longer than an hour. Now that I write that and actually check the WU information, the first result went 230 sec. As long as you have an always-on connection, it doesn't get to be too big a problem, but between 4.72's bumping the runtime for all WUs after I get one of the 12-hr ones, and the WUs that are only 20 minutes long, I run a 4-day queue for my main LHC box, just to make sure I get enough work to keep the fires lit. Me, I love the short ones--something about the old days of SETI Classic's 1 WU = 1 credit still burned into my noggin. Wish BOINC had the 'search for sweet WUs' button that SetiQueue had. 8) (j) James |
|
20)
Message boards :
Number crunching :
Whats in a name ....
(Message 10495)
Posted 29 Sep 2005 by itenginerd Post: <blockquote>I asked the Sixtrack computing guru (Eric McIntosh) to answer your question and here is his reply. -snip-<blockquote> Great information to one of the more constant puzzles of crunching for LHC. Thanks, Ben and Eric! (j) James |
Next 20
©2024 CERN