View previous topic :: View next topic |
Author |
Message |
davide405 Knight

Joined: 22 Mar 2006 Posts: 77 Location: Central USA
|
Posted: Thu May 11, 2006 9:10 pm Post subject: A Plug for Rosetta |
|
|
For those whose shrubbers work on Rosetta flavored shrubs, please consider increasing the share of time you devote to that project.
For those who don't do Rosetta at all, please consider attaching to this project.
 |
|
Back to top |
|
 |
Fart in your gen direxion I am the goatse.cx guy


Joined: 24 May 2002 Posts: 2022 Location: Regrettably for you, I'm Upwind in Upstate N.Y.
|
Posted: Thu May 11, 2006 10:13 pm Post subject: |
|
|
For a while there I was feeling like the lone ranger pimping this project .
 
 |
|
Back to top |
|
 |
Killerrabbit Major Oblivion


Joined: 23 May 2002 Posts: 4656 Location: in a rabbit hole near you!!
|
Posted: Thu May 11, 2006 11:20 pm Post subject: |
|
|
I do not wish to deal with the buggy Rosetta client.
Ni _________________
 |
|
Back to top |
|
 |
Son Goku Duke


Joined: 13 Mar 2006 Posts: 426
|
Posted: Fri May 12, 2006 12:41 am Post subject: |
|
|
Rosseta isn't quite so buggy anymore... There was one issue with 5.12 which I and some others had reported in RALPH, about it affecting the performance of our comps (well do to some debugging code it turns out), which they promptly removed for 5.13 the following day. They plan to re-introduce it for testing, after consulting with Rom Walton on making it more optimized.
Those of us on RALPH will probably find out whether it will then introduce sluggishness into our computer systems for doing other tasks or not, before it gets released to Rosseta. Rosseta has come quite a ways since RALPH branched out, and we started testing app improvements over there... |
|
Back to top |
|
 |
Mr. Snrub Prince


Joined: 20 May 2003 Posts: 1916 Location: Someplace far away...yes, that'll do.
|
Posted: Fri May 12, 2006 8:15 am Post subject: |
|
|
Oooohhh!!! My machine is running a CASP 7 target! As I understand it, CASP trials are like The All-England Summarise Proust Competition, only for protein prediction.
http://en.wikipedia.org/wiki/CASP |
|
Back to top |
|
 |
davide405 Knight

Joined: 22 Mar 2006 Posts: 77 Location: Central USA
|
Posted: Fri May 12, 2006 8:18 am Post subject: |
|
|
I echo Son Goku's experience with the Rosetta client.
There was a time, not too long ago, when I was getting errors on nearly two out of three work units. That has all been fixed.
I now get considerably better than 90% success rates.
 |
|
Back to top |
|
 |
ToeBee KWSN Castellan


Joined: 30 Jun 2004 Posts: 601 Location: Manhattan, KS
|
Posted: Fri May 12, 2006 10:09 am Post subject: |
|
|
The time it pissed me off the most is when one of their shrubs got "stuck" and just kept eating up CPU time without ever finishing. It was on a computer here at work I don't look at much during most of the month so I didn't notice it until it had already wasted a couple hundred hours of CPU time.
Anyway, that problem does seem to be corrected now and I think I saw somewhere that they actually added a check in the code that would abort the work unit if this condition was detected. Right now I'm pushing hard on LHC but then I will go back to doing a few rosetta shrubs. |
|
Back to top |
|
 |
Fart in your gen direxion I am the goatse.cx guy


Joined: 24 May 2002 Posts: 2022 Location: Regrettably for you, I'm Upwind in Upstate N.Y.
|
Posted: Fri May 12, 2006 10:19 am Post subject: |
|
|
MacG wrote: | Anyway, that problem does seem to be corrected now and I think I saw somewhere that they actually added a check in the code that would abort the work unit if this condition was detected. |
Correct, they fixed the client. I've been running it on three shrubbers for a few months. I started running it after the bad shrubs "incidents" that occurred at the end of last year and haven't had a bad shrub or hung application yet . And since then, they've installed the fix MacG speaks of. I guess before the fix they focussed on pruning the shrubbery better so no bad ones got issued, but if they do, the client now flushes them .
Mr. Snrub wrote: | Oooohhh!!! My machine is running a CASP 7 target! As I understand it, CASP trials are like The All-England Summarise Proust Competition, only for protein prediction.
http://en.wikipedia.org/wiki/CASP |
Ooooooh, most excellent .
I just checked, I'm running (seriously) a CASP6 shrub .
CASP6 was LAST year's competition ? Don't they trust me with new shrubbery ? What am I, a MORON ? Don't answer that ! Am I destined to process re-treads ? _________________
Ni ! Ni !
Flatulenty yours,
Sir Fart
The Prince of Noxious Fumigations
The Earl of Eruption
The Baron of Breaking Wind
The Marquis of the Malodorous
The Monarch of Methane
Loony Emeritus |
|
Back to top |
|
 |
Son Goku Duke


Joined: 13 Mar 2006 Posts: 426
|
Posted: Fri May 12, 2006 10:36 am Post subject: |
|
|
Yeah, they had been adding several things to deal with this problem. And one of the most recent which got tested on RALPH is the watchdog process which aborts WUs that keep re-starting before they get underway and all. That, and they have for now hard coded a max runtime into the WUs of like 1 day, so at that point the WU will end, one way or the other...
I think they're also (on Rosseta, though not on RALPH), giving credits for bum WUs. Actually, I had 1 or 2 where they seemed to have done that...
MacG wrote: | Anyway, that problem does seem to be corrected now and I think I saw somewhere that they actually added a check in the code that would abort the work unit if this condition was detected. Right now I'm pushing hard on LHC but then I will go back to doing a few rosetta shrubs. |
Yeah, lotsa people are taking a run to LHC, now that they have work. Well until our caches run out, and we can't get anymore for awhile that is. |
|
Back to top |
|
 |
buck Knight


Joined: 13 Dec 2005 Posts: 78 Location: There
|
Posted: Fri May 12, 2006 11:46 am Post subject: |
|
|
Am I'm the only one that is seeing problems of computer hangups, both BOINC and other programs after Ver 5.13 is downloaded?
 _________________ Where ever you go, There you are! |
|
Back to top |
|
 |
davide405 Knight

Joined: 22 Mar 2006 Posts: 77 Location: Central USA
|
Posted: Fri May 12, 2006 11:59 am Post subject: |
|
|
Sir Buck,
I personally have not had any trouble with v5.13, but that doesn't make your difficulties any less real.
Try going to the Rosetta website and posting on their message boards.
I have got helpful responses to most every question or concern I have ever posted, and typically in 24 hours or less. |
|
Back to top |
|
 |
davide405 Knight

Joined: 22 Mar 2006 Posts: 77 Location: Central USA
|
Posted: Fri May 12, 2006 12:01 pm Post subject: |
|
|
Sir Fart...
Don't despair! With the CASP 7 trials running until August, I'm sure you'll get a crack (poor choice of words ) at some of the new WU's.
 |
|
Back to top |
|
 |
Fart in your gen direxion I am the goatse.cx guy


Joined: 24 May 2002 Posts: 2022 Location: Regrettably for you, I'm Upwind in Upstate N.Y.
|
Posted: Fri May 12, 2006 12:04 pm Post subject: |
|
|
davide405 wrote: | Sir Fart...
Don't despair! With the CASP 7 trials running until August, I'm sure you'll get a crack (poor choice of words ) at some of the new WU's. |
I'm not so sure . I think they have "long bus" servers, and "short bus" servers. I think I'm slated to ride the short bus forever . |
|
Back to top |
|
 |
Grawlfang Prince


Joined: 18 Feb 2005 Posts: 1256 Location: The Land of our lord JC, silly walks and all
|
Posted: Fri May 12, 2006 6:39 pm Post subject: |
|
|
rosetta can't be that 'buggy', I've been scrubbing against it for months on both Linux, FreeBSD and Win2K - and I'm currently in the top 1000 users (though I've switched to QMC now so will be dropping like a stone I'm sure).
Ni!
Fang _________________ Listen; Strange women laying in ponds, distributing swords, is no basis for a system of government.
 |
|
Back to top |
|
 |
klaatuborada Yo ho! Post ho!


Joined: 08 Dec 2002 Posts: 2098 Location: by the dock by the bay...
|
Posted: Fri May 12, 2006 7:45 pm Post subject: |
|
|
I've given Rosetta prime status on my little computer. Meager shrubs, but shrubs just the same.
Here's the rub. When I first joined I did a search for the team, The Knights Who Say Ni, and the search came up with Knights Who Say Ni, so I joined it. I soon realized that it was not THE Knights Who Say Ni, so I've switched my team. I hope stats will show up soon and stat thingies will correct. (Some still say old team).
Anywho.... I'll go back to lurking.... _________________ Confucious say, "Man who lay girl on hill, not on level"
sic eas ad astra
   |
|
Back to top |
|
 |
davide405 Knight

Joined: 22 Mar 2006 Posts: 77 Location: Central USA
|
Posted: Fri May 12, 2006 8:17 pm Post subject: |
|
|
Klaatu borada nictu!
Lurk no more
There is no such thing as a "meager" shrub, or a meager shrubber.
I've got an old HP Pavillion (600 MHz CPU, a whopping 192M RAM) plugging gamely away.
 |
|
Back to top |
|
 |
klaatuborada Yo ho! Post ho!


Joined: 08 Dec 2002 Posts: 2098 Location: by the dock by the bay...
|
Posted: Fri May 12, 2006 9:10 pm Post subject: |
|
|
Ni!cto! Just figuring out how to allocate resources to different projects. I think I bit off more than I can chew because I added LHC and I think I should have stuck with just the two, (seti and rosetta), but am feeling guilty about cancelling it.
Also, just updated FRAPPR page included more BOINC and FAH info. Don't forget to add yourselves if you haven't done so already.
I think I'll go check stats boards to see if I show up now. It was the Rosetta board that really threw me, really, knights who say ni vs The Knights Who Say Ni.... pufters.... _________________ Confucious say, "Man who lay girl on hill, not on level"
sic eas ad astra
   |
|
Back to top |
|
 |
Son Goku Duke


Joined: 13 Mar 2006 Posts: 426
|
Posted: Fri May 12, 2006 9:16 pm Post subject: |
|
|
I'm not doing so much Rosseta now, outside the occasional RALPH unit to help them develop the app some more... But for the most part had switched there from predictor@home...
Even though I hadn't had many problems at all on predictor, the number of people mentioning lack of feedback, continual problems, and then nothing has me scratching my head. But the thing, I don't know the inside scoop on, but have a sneaking suspiscion on, was someone's mention of a
Quote: | well respected member of the BOINC community who got banned, for continuing to mention problems with fortran errors |
Though know names were given, this is was around the time Paul D. Buck (pretty much the creator of the whole BOINC wicki) announced his departure, and mentioned problems with certain projects, and seemed quite upset... Course, most users weren't happy to see him go, after the time and effort he put into things, which exceded many people's.
I don't know, the person mentioned there was him, however I've got a nagging suspiscion; and one I certainly don't want to go on record with, especially without anything more to go on... If it was, umm, I really have to scratch my head on what transpired there...
Despite the mentions, I've had few problems at predictor myself (though all the abeta units I last crunched turned out to be a waiste); but it's enough to leave me wanting to give my credits elsewhere. In protien folding/apps development like stuff, I now give my credits to Rosseta... |
|
Back to top |
|
 |
klaatuborada Yo ho! Post ho!


Joined: 08 Dec 2002 Posts: 2098 Location: by the dock by the bay...
|
Posted: Sat May 13, 2006 7:17 am Post subject: |
|
|
Does Rosetta need a plug because it's leaking? _________________ Confucious say, "Man who lay girl on hill, not on level"
sic eas ad astra
   |
|
Back to top |
|
 |
davide405 Knight

Joined: 22 Mar 2006 Posts: 77 Location: Central USA
|
Posted: Sat May 13, 2006 12:28 pm Post subject: |
|
|
klaatuborada wrote: | Does Rosetta need a plug because it's leaking? |
<Insert emoticon of big fat RAZZberry here > |
|
Back to top |
|
 |
|