1) Message boards : Cafe LHC : Whatever happened to LHC@home....apologies and thanks. (Message 21614)
Posted 13 Nov 2009 by Brian Silvers
Post:
Obviously LHC@home still gets support from the user base otherwise these boards would be dead for a longtime already.


What was said:


CPSS is down to less than 100 active volunteers and LHC@home support
has been dead for a year and it appears unlikely that IT will support it.



That means that the server is on, but nobody is taking care of things, as evidenced by the lack of fixing the silly reply window here and not fixing where single/double-quotes cause an escape sequence, among other trivial things. It doesn\'t sound like there is an interest, and likely the server here is just being kept running until an official \"sunset\" annoucement is made...or a server failure happens, whichever comes first...
2) Message boards : Cafe LHC : Whatever happened to LHC@home....apologies and thanks. (Message 21606)
Posted 9 Nov 2009 by Brian Silvers
Post:
Yeah... I guess IR = 0 is better than IR=2, IR=3, and far better than that \"gosh awful\" IR=5...right?

This has nothing to do with the matter at hand. Bigmac clearly stated that \"...someone with real authority and responsibility...\" implying that he does not have the authority or responsability to go ahead with LHC@home on his own.

The exerpt from the email suggests that the Management Board is not convinced at all of the power of Boinc/LHC@home and prefers an in-house solution for the calculations. Or it might be that, with the thousands of scientists on the Collider project, it is simply a story of pet projects, power games, or complete ignorance on the topic of distributed computing.


Due to the unrelenting antics of a particular individual, the admins of this server, and the only moderators of the forum that I am aware of, had to endure a constant disruptive behavior and daily babysitting of this forum just to attempt to keep a single individual from intentionally smearing the project, and thus CERN and Queen Mary by proxy.

I would think that such antics played at least some kind of part in the decision-making process on the need of utilizing the BOINC platform...
3) Message boards : Cafe LHC : Whatever happened to LHC@home....apologies and thanks. (Message 21594)
Posted 4 Nov 2009 by Brian Silvers
Post:
It looks like quite a nice cluster: http://www.theregister.co.uk/2009/11/04/lhc_xeon_endorsement/


Yeah... I guess IR = 0 is better than IR=2, IR=3, and far better than that "gosh awful" IR=5...right?
4) Message boards : Cafe LHC : Whatever happened to LHC@home....apologies and thanks. (Message 21592)
Posted 4 Nov 2009 by Brian Silvers
Post:
It seems very odd that CERN staff are heavily involved in adding VM capabilities to BOINC (see notes from the recent BOINC workshop), but LHC@home is left to rot.


My guess is that all the ruckus on the forum made the IT staff / admins / management decide we are more trouble than getting a cluster that they control and thus don\'t have to have the overhead of someone babysitting the forum.
5) Message boards : Number crunching : Exceeding maxium disk space. (Message 21256)
Posted 25 Feb 2009 by Brian Silvers
Post:
To be clear, the tasks were run on the very old sixtrack 4.67 (Windows systems anyway).
6) Message boards : Number crunching : Source code? porting, run as annonnymous? (Message 21188)
Posted 17 Feb 2009 by Brian Silvers
Post:
Due to how sensitive the results are to very minor calculation differences / errors, LHC@home do not release their code for 3rd party compiling.

And yet, they rely on computers that are most likely overclocked and running Windows. I\\\'d trust PPC over i686 any day. Oh well. I didn\\\'t really expect it to be available anyways. My RS6000 will just have to stick with SETI@HOME.


Gotta love stereotyping....

-Brian...who has had a Windows XP (32-bit) system running overclocked from 2200MHz to 2750MHz and rock solid for, oh, about 3 years now...and from 2000MHz to 2500MHz before that... (AMD 3700+ now, 3200+ then)
7) Message boards : Number crunching : Segmentation violation (Message 20592)
Posted 1 Oct 2008 by Brian Silvers
Post:
Maybe some clues will turn up.


Col. Mustard in the Ballroom with the Candlestick... ;-)



In all seriousness, has anyone considered falling back to a 5.8.x version of BOINC and seeing if the problem is still there or not?
8) Message boards : Number crunching : work , but not for my host (Message 20548)
Posted 26 Sep 2008 by Brian Silvers
Post:

I'm confused. What am I missing here?


The understanding that I have no life, for one... LOL

I'm one of several new hires where I work. The boss had a sit-down meeting with each of us new hires individually today. What he told me was that he felt that I was not shy and that I have a self-deprecating sense of humor. I laugh at myself. It may not come across here, which is why I scoffed at a certain suggestion... Text-only personas may not reflect actual personalities...

To abort redundant tasks manually one must first go to the LHC website and see which of his unstarted tasks have reached quorum then abort them in the manager. I don't see how that is just as easy as letting the software cancel the tasks for you. Furthermore, if you run a version that will comply with server requests to check in, for example every hour, then you can have redundant tasks canceled even while you are away at work or on vacation or whatever, assuming the server is configured to ask hosts to check in every hour.


Well, if I do "miss one" (or several of the shorter tasks), it is typically because I've been sleeping or am not at home. I do see the merit in doing what you're talking about, but I have seen problems with not having a network connection with various 5.10.x versions. Things have gotten better for me lately, but a while back my cable modem would drop at random times. Setting to NNT (or suspending network) would have been something I would've needed to do, which would defeat the purpose.

I'll think about it...
9) Message boards : Number crunching : work , but not for my host (Message 20545)
Posted 26 Sep 2008 by Brian Silvers
Post:
Now, from your latest post I'm getting a different picture. I could be all wrong but it sounds like maybe you are trying to make a political statement by not upgrading.


No, I just mentioned that because I do think that they have their priorities out of order and that they knew that there were older clients downloading tasks, yet either they did not test that scenario or they did and considered 0-credit as "A-OK" so long as it returned a valid result. IOW, they speak with forked tongue in regards to the value of credit parity.

Anyway, my primary reasons for not going for any newer versions are, in order:

    *The installer blowing up and leaving BOINC in a non-functioning state that took multiple reinstalls of 5.8.16 to fully correct.
    *No real need for any of the newer features.
    *The bugs of the newer versions (apppears to be sloppy testing and/or poor judgement on "go / no go" decisions to me).



Only when this project started doing the cancels did I have a \"need\", but since I do not run this project very often, and considering that when I do I run it exclusively until I've cleared out the work, then I can manually abort tasks just as easily.

10) Message boards : Number crunching : Something is wrong managed!!! (Message 20542)
Posted 26 Sep 2008 by Brian Silvers
Post:
If hosts would contact the server more often then more redundant tasks would get canceled. The project managers can direct hosts to contact the server more frequently but they are not doing so.


A question in regards to this:

Would setting a project to "No New Tasks" (or whatever it is called in newer versions) end up making it to where this would not work?

Reason I'm asking is because I am pretty sure that if I do that (set to NNT) with 5.8.16, if I have a pending scheduler connect on a countdown, BOINC won\'t even attempt to connect when the countdown is over. It is either that, or it does the connect that time, but no more....
11) Message boards : Number crunching : work , but not for my host (Message 20541)
Posted 26 Sep 2008 by Brian Silvers
Post:
Anyway, I noticed after posting that the status box listed the project as \\\"down\\\", but with work available. I didn\\\'t have the time to go back to clarify...


Actually the scheduler was down but everything else was running. There were several thousands results available but if the scheduler was down then the results won\\\'t get sent when hosts request work.


When I looked, the only things running were the data-driven web pages and the upload/download server, i.e. the top two entries on the server status page. Everything else had a status of \"not running\". This is getting into minutae though... Since we (you, me, and littlebouncer) all looked at the situation at different times, the conditions most likely were different for all 3 of us, even if only slightly.

Bottom line is, if the scheduler is down, nobody is getting work.


I recall you saying you tried to install 5.1.x but it did not go well, crashed or something. I vaguely recall one release in the early 5.x series which was quickly withdrawn because it caused a lot of grief. Perhaps you were unlucky enough to try that one.


It was 5.10.28 that blew up on me.

I tell you what. When David Anderson and/or Eric Korpela make 5.10.45 the minimum version that will get work from SETI instead of allowing 3.x and 4.x clients to continue to download work, then I\'ll consider it. The \"fear\" of not forcing a newer version is that people were behind NTLM proxies and they would be \"cut off\" and that those people MIGHT be sponsors and that they just absolutely could not \"cut off\" a sponsor, nor could they even ask any of the sponsors if they were having that problem or if they knew people who did have problems, nor could they contact a sponsor and ask them to update. Meanwhile, knowing all this, they did not do testing of the newest science applications there under those versions of BOINC.

That\'s what I call \"shameful\"...
12) Message boards : Number crunching : work , but not for my host (Message 20533)
Posted 25 Sep 2008 by Brian Silvers
Post:

I think LHC@home does not use HR. I have primarily Linux hosts. I have not received LHC work for quite some time but if I recall correctly they were almost always grouped with Windows hosts so heterogenous redundancy seems to be in effect here.


That is the case. My line of thinking this moring was related to being involved in a Homogeneous Redunancy project that currently has a very severe problem with HR (Cosmology). I was in a bit of a rush, and have something else on my mind too (death of someone I grew up with), so I was a bit...terse (and perhaps tense) this morning...

Anyway, I noticed after posting that the status box listed the project as \"down\", but with work available. I didn\'t have the time to go back to clarify...

As for the newer versions of BOINC, I\'m still not convinced to try even 5.10.45, let alone 6.x.x. Many things are for Vista, which I don\'t have and don\'t plan to have anytime soon. As for the non-Vista stuff, most of that seems to be geared towards making the messages more palitable for \"non-technical\" users that might stumble across them. I think it also has some fix for NTLM proxies, but since I don\'t use one, that\'s not an issue I need fixed.
13) Message boards : Number crunching : work , but not for my host (Message 20530)
Posted 25 Sep 2008 by Brian Silvers
Post:
To clarify, if there was work available for other hosts, you will be told that there was work available for other platforms in the response message from the scheduler.


Close, except littleBouncer\\\'s hosts are Intel and AMD processors running on WinXP. A non-supported platform is not the issue here. The thread title is a tad misleading for the uninitiated.


I wasn\'t talking about a user of a non-supported OS platform, but was thinking more about the issue that Cosmology is having with their HR implementation. Since LHC does not use HR (to my knowledge), then I\'ll agree that I was using either a poor example or did not explain myself enough. I quickly typed out that message earlier as I was waiting for my turn in the bathroom so I could finish getting ready for work...

As for the sched_op_debug thing, I thought that it would still tell you the HR-related messages, but since I do not use the newer versions of BOINC, I could very well be mistaken.
14) Message boards : Number crunching : work , but not for my host (Message 20523)
Posted 25 Sep 2008 by Brian Silvers
Post:
To clarify, if there was work available for other hosts, you will be told that there was work available for other platforms in the response message from the scheduler. There are tasks \"in progress\" right now, but no tasks needing to be sent out. This project is at least fairly reliable when it comes to sending correct messages to BOINC hosts when there is no work, unlike the major problems being encountered over at Cosmology.

Also, bear in mind that the actual LHC device is down until sometime next year, so there may not be any work available for a few months.
15) Message boards : Number crunching : Something is wrong managed!!! (Message 20490)
Posted 22 Sep 2008 by Brian Silvers
Post:
littleBouncer's plan is not progress, it is regression to the old, slow, wasteful way.


I think the walls of some structure somewhere just collapsed, because we agree that the "plan" put forth by littlebouncer defeats the purpose of not processing tasks that don't need to be processed...

FYI, if you look at my results list, I've been manually aborting tasks that already met quorum. I have one task to process that I may end up missing reporting by a couple of hours. Not sure yet. I also have 2 "ghost" tasks that I don't know what happened to. I haven't checked to see if they were downloaded or not, but they are definitely not on the box waiting to be processed (and they have met quorum too)...
16) Message boards : Number crunching : BOINC 5.10.x?? (Message 20046)
Posted 9 Sep 2008 by Brian Silvers
Post:
Down=versioning isn\\\\\\\'t really an option. I have other work running (long jobs) and the change from Version 6.x.x back to Version 5.x.x will likely trash all the work already done. I have taken this route in the past and seen it happen before.


I went from 6.2.xx to 5.10.xx and it worked fine. It took a little more time but I didn\\\'t loose my cpdn @ 86% complete and 749hrs.

EDIT: Don\'t forget to set U\'r manager to SNOOZE before you start the downgrade!


I stick with 5.8.16 for a reason... ;-)

Maybe Mr. Anderson should try fixing his software before worrying about his other pet peeves?
17) Message boards : Number crunching : Can't display WU (Message 19987)
Posted 2 Sep 2008 by Brian Silvers
Post:
:bump:

Viewing workunits still give a PHP error.

Also, the repair_forums script needs to be run to repair the escaping of ' and " marks...


Hmmm... It appears as though someone took care of the PHP error...

Thanks!

Now, if only the repair_forums script could be run and the issue with the tiny message compose area solved....
18) Message boards : Number crunching : Can't display WU (Message 19981)
Posted 2 Sep 2008 by Brian Silvers
Post:
:bump:

Viewing workunits still give a PHP error.

Also, the repair_forums script needs to be run to repair the escaping of ' and " marks...
19) Message boards : Number crunching : Can't display WU (Message 19927)
Posted 13 Aug 2008 by Brian Silvers
Post:
I've noticed the same same problem, Viking69. You will also find slash marks in some posts. there are two workarounds necessary: copy and paste into a blank text editor (notepad) to see the entire post, and then use the edit feature to eliminate the slash marks.

Good luck,
Kibble


There's a fix for that...


I checked in some code (in html/ops/repair_forums.php)
to change " to " and ' to ' in:
- forum posts
- forum thread titles
- profile text

I've run this on SETI@home and the BOINC message boards;
I recommend that you run it to make your message boards look nice.

-- David


Perhaps it would also fix the annoyingly tiny message window as well...?
20) Message boards : Number crunching : Federal suit filed to stop... (Message 19319)
Posted 29 Mar 2008 by Brian Silvers
Post:
Federal law suit

Just for information.


You've heard of NIMBY as a reason for not doing something, now you have....

NOTEP (Not On The Entire Planet)
or
NOPE (Not On Planet Earth)


Next 20


©2021 CERN