British (Mar 17 2009)

Message boards : Technical News : British (Mar 17 2009)
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Matt Lebofsky
Volunteer moderator
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 1 Mar 99
Posts: 1444
Credit: 957,058
RAC: 0
United States
Message 876616 - Posted: 17 Mar 2009, 21:37:38 UTC

Hello again. Sorry about missing a couple days there. The end of last week I did write a tech news item that I neglected to post as I got suddenly very busy at the end of the day with random programming tasks, and yesterday I was lost in many meetings and other post-weekend catchup. So be it. Here I am now.

The end of last week I was a stand-still with various projects, so I chipped away at neglected chores and other nagging annoyances. Like our new mail server's log filling up with cryptic automounter messages regarding a machine we haven't had on line in five years - I finally tracked this down to Eric's home-grown spam challenge script which made reference to this machine in its LD_LIBRARY_PATH. I also tried and failed to figure out why one of our systems, configured exactly like the others, refuses to acknowledge the lab-wide legato backup server. And I cleaned my keyboard for the first time ever (which was gross after years of eating at my desk, and this was probably not helping the lingering ant problem). Then I got lost in NTPCker stub web page design.

Yesterday there was much discussion about radar. Dan, recently back from Arecibo, confirmed some things and had news about others. The radar blanking code I took over and improved upon had faulty logic, caused by some early misunderstandings (not mine) about how the radar behaved. Most of the radar we see is from the airport, and that's all the hardware blanker thwarts. However, there are 5 other patterns we detect, including the aerostat balloon radar. So one problem is that at times we're seeing a jumble of various radars, making it very difficult to "lock on" and blank them. I'm working on that now. One other point is that the radar frequencies are all pretty much out of our band (typically around 1.3GHz - we're looking around 1.42GHz), but nevertheless are so loud they jam our receivers. However, sometimes if certain projects call for it the Arecibo operators turn on a high pass filter so that the radar frequencies under ~1.4GHz are completely silent. When this happens (about 20-25% of the time) our data are incredibly clean, even without hardware blanking. Of course, since we're piggybacking we can't control when the filter is on, but we do keep track of it in our data headers. We might prioritize this cleaner data for astropulse, which is far more sensitive to radar than SETI@home.

Today had the usual outage for mysql database backup/compression. I took extra time while everything was quiet to move a lot of big files around the raw data storage server - that's mostly why we were slow to get out of the outage this time around, but at least now I can start emptying the latest shipment of drives from Arecibo. Speaking of drives, there was some discussion about that, too. We may start trying to partially send data over the net, if not completely. We thought this was impossible due to bandwidth constraints, but operators at Arecibo told us to give it a shot. This is low priority since, however annoying, the drives, their enclosures, and the shipping rigamarole works well enough right now.

In general the public-facing servers continue to behave themselves. It's been a good couple of weeks. I don't believe in jinxes so I don't mind saying as much. I will say that the workunit storage server is filling up again - a factor of astropulse actually performing well, and workunits sitting around a long time waiting to validate. If it does fill up we'll have to deal with it.

- Matt

-- BOINC/SETI@home network/web/science/development person
-- "Any idiot can have a good idea. What is hard is to do it." - Jeanne-Claude
ID: 876616 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 876647 - Posted: 17 Mar 2009, 22:02:13 UTC

Matt,
tried to send in 4 results and got this message.... 3/17/2009 5:55:24 PM SETI@home Scheduler request failed: Failure when receiving data from the peer

Is that something on your end? I haven't seen that message before. When I tried again I got this message... 3/17/2009 5:58:09 PM SETI@home Scheduler request failed: Server returned nothing (no headers, no data)




PROUD MEMBER OF Team Starfire World BOINC
ID: 876647 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 876665 - Posted: 17 Mar 2009, 22:08:18 UTC

Re the thread title - here in British-land, it's still March 17.....

.....and that means the title should be Irish, because it's St. Patrick's day. The Irish don't like the British very much, because of little accidents like the Famine (potato blight). Without which, you wouldn't have had so many navvies to build your railroads. It's a rum old world.
ID: 876665 · Report as offensive
Richard Haselgrove Project Donor
Volunteer tester

Send message
Joined: 4 Jul 99
Posts: 14649
Credit: 200,643,578
RAC: 874
United Kingdom
Message 876682 - Posted: 17 Mar 2009, 22:30:13 UTC - in response to Message 876616.  

I will say that the workunit storage server is filling up again - a factor of astropulse actually performing well, and workunits sitting around a long time waiting to validate. If it does fill up we'll have to deal with it.

We've been musing about this in Number Crunching, specifically in MB, AP, and Cuda requests.

Many people have been commenting that they are seeing an unexpectedly high proportion of AP_v5 in their downloads, and would like more MB. We also noted that the AP splitters are working through the 'tapes' on the Server Status page far faster than the MB splitters, to the extent that there are 31 '(done)' AP tapes waiting to be split for MB.

We are surmising that this is a side-effect of the change you made at The End of All Things (Oct 30 2008) - activating the "-allapps" flag.

Since then, of course, we have had the deployment of the CUDA app for MB, which is also starting to run smoothly. So the demand for MB tasks from the faster hosts is probably leaving a lot of AP tasks for slower hosts - the opposite of the desired distribution.

Would it be possible to skew the priorities in "-allapps", to feed (and hence split) a higher proportion of MB? Joe Segur suggests (in the thread) that there are "weightings" available in the server code to achieve this re-balancing.
ID: 876682 · Report as offensive
Urs Echternacht
Volunteer tester
Avatar

Send message
Joined: 15 May 99
Posts: 692
Credit: 135,197,781
RAC: 211
Germany
Message 876687 - Posted: 17 Mar 2009, 22:41:45 UTC

British (May 17 2009)


Must be a 2 month sized rift in the space-time-continuum between the so called "new world" and the so call "old world". Didn't know that our good old europe is so far back.

Matt, try to get some baking powder and scatter it on the ants streets (they usually use the same (smell) pathes again and again). The powder should make them loose the smell and they get disorientated. After a while they give up and search at some other place for food.
_\|/_
U r s
ID: 876687 · Report as offensive
Profile Matt Lebofsky
Volunteer moderator
Project administrator
Project developer
Project scientist
Avatar

Send message
Joined: 1 Mar 99
Posts: 1444
Credit: 957,058
RAC: 0
United States
Message 876689 - Posted: 17 Mar 2009, 22:46:38 UTC
Last modified: 17 Mar 2009, 22:47:06 UTC

The thread title (which was originally slated to be used last thursday but I never got around to posting) is simply keeping in line with this month's theme. The May part is a bug - I'll fix.

- Matt
-- BOINC/SETI@home network/web/science/development person
-- "Any idiot can have a good idea. What is hard is to do it." - Jeanne-Claude
ID: 876689 · Report as offensive
Profile Fred J. Verster
Volunteer tester
Avatar

Send message
Joined: 21 Apr 04
Posts: 3252
Credit: 31,903,643
RAC: 0
Netherlands
Message 876700 - Posted: 17 Mar 2009, 23:16:26 UTC - in response to Message 876689.  
Last modified: 17 Mar 2009, 23:25:41 UTC

Hi Matt, anyway, thanks for the weekly update, the WU handling appeares less faulty as in some other weekends.
A little bit[OFF TOPIC]
I think, the latest BOINC versions from 6.2.19 and up, have, except for (better) CUDA handling, also better workload handling.
No extreme amounts of MB WU's, AP WU's on too slow hosts(as far as I see on my hosts. ;^)).
Also not being out of work, since using the latest BOINC versions, also not being overloaded with WU's.
Even the task managments, which didn't work too good with 5 series BOINC, works better and devided tasks bettter .

And, as far as I know, NO major problems, now, that's always good news.
ID: 876700 · Report as offensive
Profile Neil Blaikie
Volunteer tester
Avatar

Send message
Joined: 17 May 99
Posts: 143
Credit: 6,652,341
RAC: 0
Canada
Message 876757 - Posted: 18 Mar 2009, 1:47:10 UTC - in response to Message 876682.  
Last modified: 18 Mar 2009, 1:48:16 UTC

I think I have finally sorted my "System clock was turned backwards; clearing timeouts" problem out, only to find that requesting my 4 day cache, I get 3 AP v5 5.03 work units and 0 MB.

Hopefully something could be done about this.

Thank you though Matt for taking time out to post your tech news. No harm in not posting, work is work and should take priority over telling us crunchers what is going on behind the scenes.

Seem you guys have achieved a lot today and should be provided with seti cruncher approved beer for your consumption off site at your leisure. :-) or on your newly cleaned desk!
ID: 876757 · Report as offensive
Profile Allie in Vancouver
Volunteer tester
Avatar

Send message
Joined: 16 Mar 07
Posts: 3949
Credit: 1,604,668
RAC: 0
Canada
Message 876815 - Posted: 18 Mar 2009, 5:55:11 UTC - in response to Message 876616.  
Last modified: 18 Mar 2009, 6:01:12 UTC

but nevertheless are so loud they jam our receivers.



Stupid-girl suggestion: rather than biasing the system against frequencies, bias it against signal strength. That is, any signal above x dB has to be terrestrial and thus, can be ignored.

{edit} I realize that it is a lot more complex than that, but if the first level of filtering is signal stength, then a lot of the human created crap would be gone in an instant.
Pure mathematics is, in its way, the poetry of logical ideas.

Albert Einstein
ID: 876815 · Report as offensive
Norwich Gadfly
Avatar

Send message
Joined: 29 Dec 08
Posts: 100
Credit: 488,414
RAC: 0
United Kingdom
Message 876836 - Posted: 18 Mar 2009, 9:25:49 UTC - in response to Message 876682.  

I have my preferences set to any type of work, and recently the only S@H units I've been crunching only these Astroturf units, which I'm quite happy about. I have a 4 core PC and I split the spare cycles between this project and the World Community project, so if ET isn't found at least some tangible benefit should ensue.

I enjoy reading these technical reports, and am very impressed by the way Matt and the rest of the team are running this project. You are doing very well, carry on everybody !
ID: 876836 · Report as offensive
Profile Dr. C.E.T.I.
Avatar

Send message
Joined: 29 Feb 00
Posts: 16019
Credit: 794,685
RAC: 0
United States
Message 877174 - Posted: 19 Mar 2009, 12:16:31 UTC






<snip> prioritize this cleaner data for astropulse



< nice work there Matt - . . . 'like' the suggestion [quote above] - Lunatics workin' real well btw . . .

Accolades to each of you @ Berkeley . . .


BOINC Wiki . . .

Science Status Page . . .
ID: 877174 · Report as offensive
Profile Fred J. Verster
Volunteer tester
Avatar

Send message
Joined: 21 Apr 04
Posts: 3252
Credit: 31,903,643
RAC: 0
Netherlands
Message 877177 - Posted: 19 Mar 2009, 12:45:58 UTC - in response to Message 877174.  
Last modified: 19 Mar 2009, 13:38:28 UTC

Hi, Matt, keep up the good work!
I noticed a certain 'dislike' for AP tasks. I've looked at a few 'wu_3 or 4', most off them had just been aborted!?
All these 'abortions', mostly comes from slower hosts, < 2500 MFLOPS/core and NO optimized processing.

Are AP units faster to split, because the files are larger? And how much computing time goes in radar-blanking these units, if it's done before they are send out, stupid question and probably (very) time-consuming?

If the radar-frequency is known, isn't it possible, to filter this frequency to below the detection threshold and the harmonics, too?
Or probably already done? Although a useable broadband filter isn't easy to imply. Mostly takes more amplification for the outcoming signal, if possible, without adding to much distorsion or more harmonics?!

And the CUDA handling of MB WU's, is much better, so I'm NOT surprised with this result, but since Astro Pulse IS a new way to search, they DO have to be computed.

So why, not pick some fast enough host's, for crunching these.
F.i. I have 2 Q6600's, they do them in 12* -16 hours and 1 QX9650, which does 4 AP WU's in 10 hours!

*(@ 3GHz.)
Maybe, AP handling, isn't quite clear for every one, took me some time, but I'm gettin 'old'
But, I DO like the adjusted credit, 1230 Cobblestones (x4)in 10 hours, on that one.
Applies to (V5.0 &) V5.03 (optimized, that is)
So you may send me lots of them, ;^)
ID: 877177 · Report as offensive

Message boards : Technical News : British (Mar 17 2009)


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.