LHC@home

Well, it takes a plain text editor and about 15 seconds to clear the LTD for a given project. Still, I entirely do not see the point of doing so. The only tinkering I did in BOINC was to disabled work for all projects so that CPDN would finish a WU in time (that version of the app reset its percent done for each of 3 phases, meaning that until phase three, BOINC didn't have a freakin' CLUE how much longer it would really take to complete the unit). Aside from that, I've let BOINC manage its own projects as it sees fit. I absolutely do not see how letting BOINC poll LHC could cause somebody to lose 75% of their computing power.

Not saying it cannot happen, but I've never seen LHC causing trouble on any of my Linux or Windows boxes, and I've been around since the beta days. YMMV, I suppose.

---

Yeah, those tricks will also work, but I'll have a hard time manually editing those files two times a day on 24 Systems (Debtviewer is not Network capable).

All in all, way too painful just to 'hang on' onto a Project that hasn't any work anyway.
(and in the first place, none of this should be needed but maybe that's just me)

If the overall Situation doesn't change by end of next month, LHC will finally have to make room for an active Project.

---

Scientific Network : 45000 MHz - 77824 MB - 1970 GB

(and in the first place, none of this should be needed but maybe that's just me)

IMO, none of it is needed. While LHC has no work, it has no negative impact on the performance of your other projects; all it does is generate some noise in the logs and generate some network traffic. When LHC does release some work, your systems will crunch what they can yet not cause other projects to fail on their deadlines.

So, where's the problem? It's an honest question. I'm not trolling.

---

@FalconFly : Don't panic, you sure will need a nice supply of lasagna for this project sooner or later :-)

p.s.: the newsletter says there is progress, it's just taking place in other parts of the project now, not on our prediction side.

(and in the first place, none of this should be needed but maybe that's just me)

IMO, none of it is needed. While LHC has no work, it has no negative impact on the performance of your other projects; all it does is generate some noise in the logs and generate some network traffic. When LHC does release some work, your systems will crunch what they can yet not cause other projects to fail on their deadlines.

So, where's the problem? It's an honest question. I'm not trolling.

There is one real problem, some versions of the scheduler will end up not requesting work from other projects. The LTD can build up on the project without work, especially if there is a long queue. If the other projects get into deadline trouble while the project without work is not deffered then the project without work has it's LTD adjusted. Once the LTD gets low enough (on the active project) the client will not try to get work from the project until a CPU is starved. Keeping a short queue will delay or prevent this problem.

---

BOINC WIKI

BOINCing since 2002/12/8

Basically the Queue will be reduced downto 0 (Zero), eventually requiring 24/7 online Connection in order not to starve all Systems.

With LTD easily exceeding two Million Seconds, the entire Scheduling and Cache System is basically becoming corrupted.

LTD also has effect across all attached Projects, even e.g. when resetting Projects will lead to insane LTD quirrels as the Scheduler re-assigns the cumulative LTD across all attached Projects (in plain : resetting LHC at 2M LTD will add some 500k LTD to 2 other attached Projects, sending them into deep trouble, as this can't simply be compensated in a few hours crunching)

Having no 24/7 online connection and a mandatory Cache size of at least 1.5 days to cover my standard offline periods, having LHC attached without setting it to "Suspended" will quickly result in severe loss of CPU power (I've seen my entire Network run dry after as little as 30 Minutes, which normally holds some cumulative ~1000 hours CPU time worth of work to cover 1.5 days) and is overall just painful.
Same problem case arises, when unattaching from such a Project; the LTD Delta will be re-assigned across the active Projects, sending their (until then normal ) LTD skyrocketing.
And that's not what I keep a ~$20k Network running for...

So far, I'm not aware of a newer BOINC Version that fixes this bug, that would make life with LHC alot easier.

---

Scientific Network : 45000 MHz - 77824 MB - 1970 GB

Actually the newest alpha test version is suposed to work much better in this situation. It is supposed to always keep at least your min queue amount of work on hand even if your debts are severely out of whack. I am not really sure if it works or not. Even though I consider my current queue excessively large (0.8 days) it still is normally overfilled just because I am attached to so many projects.

---

BOINC WIKI

BOINCing since 2002/12/8

FalconFly:

Having no 24/7 online connection and a mandatory Cache size of at least 1.5 days to cover my standard offline periods, having LHC attached without setting it to "Suspended" will quickly result in severe loss of CPU power (I've seen my entire Network run dry after as little as 30 Minutes, which normally holds some cumulative ~1000 hours CPU time worth of work to cover 1.5 days) and is overall just painful.

Do you test 5.5.16? I think you would be a valuable tester of the new work fetch code.

Norbert

---

No, I don't test because in case of Problems I'd have way too much work to cope with.
I'm very dependent on stable production code, everything else is nothing for me.

I used to file bug reports in the early days of BOINC but as they were either never acknowledged or implemented halfheartedly until a full year later or so, I stopped investing time into that.

---

Scientific Network : 45000 MHz - 77824 MB - 1970 GB

Ok - I've had enough of this.

I made a single post wishing you all well, and gave as my reason for departing as the poor state of the project.

Since then I seem to have been turned into the devil incarnate beacuse I 'whinge and moan', because I 'post a dummy spit', because I 'hate the project', blah, blah, blah...

my conclusion is not the same as this naked person.
but let us give him credit, he has a point. there is no information.

we all know workunits are only sometimes available.
but information should be allways available.

with timely information this cruncher could have been kept onboard.

information on the site every 2 or 3 months is not the way to keep your crunchers. it should at least be weekly.

(bas)

---

---

me@rescam.org

Just a couple more weeks and then my plan to put you all to sleep so I can load ALL the LHC data will begin!


Until then I will work on Einstein and some beta-seti just for fun.










But then if you do wake up just join my...............

---

Volunteer Mad Scientist For Life

It would be desirable to hear managers of the project.
When it is possible to expect the beginning of work. Or the project is completed?

---

BOINC.RU - BOINC in Russia "Russia Team"