Posts by ivan


log in
1) Message boards : Number crunching : Bad News on BOINC funding (Message 1699110)
Posted 15 hours ago by Profile ivan
Hmm, I wonder if all this is behind the lethargy that's overtaken CMS@Home-dev lately? I'm starting to seriously reconsider my position as its public face.
Interesting times, I'd always felt bending our workflows to fit with the BOINC model was a bit Procrustean, be interesting to see if this prompts a re-think.
2) Message boards : Number crunching : Two Trillion events reached by the Test4Theory BOINC-CernVM project !!! (Message 1697901)
Posted 4 days ago by Profile ivan
Ivan you should have posted this one mate in the sience section , CERN wahoooo go CERN

Yeah, maybe, tho' I was looking for a response from the crunchers. :-)
Just had a two-day meeting of the UK institutes in our experiment; heard a number of interesting items I wasn't aware of before. This could be an interesting year for LHC!
3) Message boards : Number crunching : 'rm -rf *' considered harmful (Message 1697896)
Posted 4 days ago by Profile ivan
It's fine. There is no rush. It leaves a bit of work for the rest of us. :)
Cheers, mate, shall I leave it until Monday then?
4) Message boards : Number crunching : 'rm -rf *' considered harmful (Message 1697871)
Posted 4 days ago by Profile ivan
I can't actually crunch on the machine at present, the server-room A/C failed yesterday so I shut down all S@H jobs so we can do our "work" work without reaching harmful temperatures.

I met a service tech in the lift on my way out tonight -- he'd just replaced the faulty thermistor in the A/C unit, so I'll see if I can start bringing up background S@H jobs on all the servers again tomorrow (I have been running just on one that's designated as my "personal" development machine), without overloading the system. My RAC is starting to take a beating...
5) Message boards : Number crunching : Suddenly BOINC Decides to Abandon 71 APs...WTH? (Message 1697126)
Posted 6 days ago by Profile ivan
Note Ivan's tale of woe in 'rm -rf *' considered harmful, especially the [edit] to the opening post.

In that case, deleting all (or a substantial proportion) of the files was equivalent to a true 'detach', so the action taken was appropriate and indeed desirable. We need to be sure we keep the handler for that situation.

I'm glad someone found some good in the situation. :-) Me, I'm off-campus for two days, will revisit the problems on Friday.
6) Message boards : Number crunching : 'rm -rf *' considered harmful (Message 1697118)
Posted 6 days ago by Profile ivan
How blue was the sky today, at 30.4 C outside (and just a few tenths less in my office...)?
7) Message boards : Number crunching : 'rm -rf *' considered harmful (Message 1697107)
Posted 6 days ago by Profile ivan
Apologies to wingmen on my older Xeon/CUDA server. In a fit of heat- or svn-frustration-induced fever today I typed the fateful command 'rm -rf *' thinking that I was connected to another computer in another window in another workspace and not in my $HOME directory. :-( By the time I realised that the command was taking longer than it should, and hit Ctrl-C, I'd lost all files (alphabetically) down to boinc* -- which included BOINC, so all my WUs are lost.

Sorry.

boinc_repository was still intact, so I've rebuilt the boinc executables, and the CUDA executable is still in its build directory; so I just need to recreate the app scripts. Now we're up again after maintenance I'll try to reattach, merge with the old instance, and abort/reset so the WUs can go to new homes. I can't actually crunch on the machine at present, the server-room A/C failed yesterday so I shut down all S@H jobs so we can do our "work" work without reaching harmful temperatures.

[Edit] Ah, it looks like the server automagically merged the new instance and abandoned all the pending tasks. Now to wait for the A/C to be repaired before starting again. [/Edit]
8) Message boards : Number crunching : The kittyman has fixed the rig, work on the way, wingys! (Message 1696475)
Posted 8 days ago by Profile ivan
Dunno.

As long as you're back on the air, that's what matters. I must find the (round) tuits to look at one of my home servers that died last month. It's a bit old in the tooth (Core2Quad) and worth less than 2,000 RAC but it is my disk server (tho' I do have a NAS as backup to it).
9) Message boards : Number crunching : Two Trillion events reached by the Test4Theory BOINC-CernVM project !!! (Message 1696432)
Posted 9 days ago by Profile ivan
Just received this message. Congratulations to all involved!

Date: Sun, 28 Jun 2015 06:40:57 +0100 From: Ben Segal <xxx@cern.ch> Subject: Two Trillion events reached by the Test4Theory BOINC-CernVM project !!! Dear all, Today we reached the total of 2 Trillion events simulated since 2011 by Test4Theory volunteers for the CERN PH-TH project MCPlots. We just announced this on the Test4Theory project News forum, and thanked our volunteers, but maybe a wider announcement would be in order? For example, back in September 2013 when we passed the 1 trillion mark, a lot of tweeting was done and CERN published an excellent article about the project: http://home.web.cern.ch/about/updates/2013/09/crowd-sourced-computing-platfo rm-reaches-one-trillion-events . And since then most of the LHC experiments have started BOINC-CernVM projects. Anyway, please feel free to forward this message as appropriate. Best wishes and thanks to all those who have helped to achieve this result, Ben and the Test4Theory team
10) Message boards : Technical News : Mid June Update (Jun 23 2015) (Message 1694900)
Posted 13 days ago by Profile ivan
Thanks for the info, Matt. Getting some cricket graphs back up would help us armchair quarterbacks second-guessing what's happening at any given instant. :-)
11) Message boards : Number crunching : Furtherst north sight for the telescopes (once in a year - repeatedly) (Message 1693620)
Posted 17 days ago by Profile ivan
A midsummer's night dream?
Tullio

Night dream right. It's so much light.

Yeah, I spent a day^H^H^Hweek in Trondheim one summer. ...and took a 3-week cruise from Mawson to Melbourne starting about December 7th another year.
12) Message boards : Number crunching : Panic Mode On (98) Server Problems? (Message 1692545)
Posted 20 days ago by Profile ivan
Strange...
I have a strange feeling of emptiness.

Mmm. Came in this morning to find my wall of gkrellm monitors largely black. Well, the 12-core is down at the moment; I tried finally to put it and the dual-20-core servers into a rack yesterday, but it seems when I rebuilt the rack I put the uprights fractionally closer than the standard rack dimensions, and the 10C's rack rails are about 5mm too long, fully collapsed, to fit.
But one 20C's display is blank, and the other shows only three jobs running.
After a bit of panicking, I eventually realised the fault was in Berkeley, not North Heathrow. Should be a trifle quieter in the server room today.
13) Message boards : Number crunching : Current motherboards that can support 6+ GPUs? (Message 1692273)
Posted 20 days ago by Profile ivan
http://www.supermicro.co.uk/products/system/4U/4028/SYS-4028GR-TR.cfm
I'm expecting a quotation on something not quite that big soon. Chap in the department had an Intel box for evaluation a couple of years ago -- it had a solid row of dual-width PCI-e slots across the back of the unit.
I'm...

I'm...

I can live in my office if I sell my house, right?

~W

I'm sure it's been done...
14) Message boards : Number crunching : Current motherboards that can support 6+ GPUs? (Message 1692093)
Posted 21 days ago by Profile ivan
http://www.supermicro.co.uk/products/nfo/gpu.cfm
15) Message boards : Number crunching : Panic Mode On (98) Server Problems? (Message 1691543)
Posted 22 days ago by Profile ivan
*sigh* Oh well, then. Lesson learned. I don't know what the max value for <save_stats_days> is... but 36500 is apparently too much. 3650 works, so for now, I'm going to try 5000 and see if that works. Let that be a lesson to all of you, in case any of you were wondering about that, or thinking of extending your stats out a bit longer. You can absolutely for sure do 10 years... but I don't know how much farther you can go on that.

Sounds suspiciously like it might be 32767 for some reason -- max value for a signed 16-bit integer.

In use, the days get multiplied by 86400 and the result needs to still be a positive 32 bit integer. So it seems 2^31/86400 = 24855 would be the maximum.

Ah, yes, that sounds more plausible. Thanks Joe, I was hoping S@H wasn't still running on 16-bit PDP-11s!
16) Message boards : Number crunching : Panic Mode On (98) Server Problems? (Message 1691203)
Posted 22 days ago by Profile ivan
*sigh* Oh well, then. Lesson learned. I don't know what the max value for <save_stats_days> is... but 36500 is apparently too much. 3650 works, so for now, I'm going to try 5000 and see if that works. Let that be a lesson to all of you, in case any of you were wondering about that, or thinking of extending your stats out a bit longer. You can absolutely for sure do 10 years... but I don't know how much farther you can go on that.

Sounds suspiciously like it might be 32767 for some reason -- max value for a signed 16-bit integer.
17) Message boards : Number crunching : Bitcoin Utopia Fund Raiser (Message 1690572)
Posted 24 days ago by Profile ivan
Yes, that's about it. Another way of describing the UK system is to say that if you want the charity to get 1,000, you only have to write a check for 800.

Richard, your spell-chequer is borked!
18) Message boards : Number crunching : SETI/BOINC Milestones [ v2.0 ] - XXIX (Message 1689876)
Posted 26 days ago by Profile ivan
ivan, I just don't even like to think what the hell you used to get to where you are. It makes me do things like think how my wife would react if I re-mortgaged the house to buy a rack full of Titan X...

Well, you've got the same RAC as me, but a few more computers :-) Time was I always had faster/better computers at home than at work, but that's changed. At least I'm allowed to run "stress testing" at low priority, it's all Science after all. That pair of 20-core machines I got recently help, but with a 100 WU limit, they don't hold much of a cache!
19) Message boards : Number crunching : SETI/BOINC Milestones [ v2.0 ] - XXIX (Message 1689537)
Posted 27 days ago by Profile ivan
Graham Middleton, I have you firmly in my sights. According to boincstats I'm only about a month from overtaking you to 21st place in the UK. :-)
Well as I said to the meet last week, I'd just slipped up another place overall -- in less than a month I'll be up another.
And from there? WORLD DOMIN...
Sorry, was I doing megalomania again? I do apologise!

My UK domination is under threat from my friends at Imperial College, "keeping their server room warm" by running S@H while there's not enough LHC jobs to crunch. Thankfully we started 13 TeV running last week, so maybe they'll get back to HEP stuff soon. Otherwise they'll overtake me around August. I must speak sternly to them at our meeting early next month -- they should be doing CMS@Home(pre-beta) instead!
20) Message boards : Cafe SETI : Seti Meet in London (Message 1687509)
Posted 4 Jun 2015 by Profile ivan
Just to report that the mini-meet in London went very well with 7 people being there.

Who were they Chris?

I don't remember.
I think it was me, Anniet, Chris S and Graham Middleton down one side of the table; Uli, Woodgie and Havoc down the other -- but I may have got the aliases wrong (problem with long-distance high-speed motorcycling sans earplugs -- you lose high frequency hearing and catching conversation in a noisy background is difficult). (I thought there were eight at one point, but that may have just been the £5.xx/pint Punk IPA -- I went around to Clerkenwell Green later and in The Crown local lager was 'only' £5.00 :-). <sheesh!>
Bernie is going to get a lot of spam in my name as I used his mobile phone number to verify my usage of the "free" WiFi...


Next 20

Copyright © 2015 University of California