Catchup (Nov 09 2015)

Message boards : Technical News : Catchup (Nov 09 2015)
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 1740982 - Posted: 10 Nov 2015, 0:29:29 UTC

Okay. Every time I put off writing a tech news item a bunch more stuff happens that causes me to continue putting off even further! So here's a quick stream-of-consciousness update, though I'm sure I'm missing some key bits.

First off, the AP migration is officially finished! As a recap, things got corrupted in the database late last year - and to uncorrupt it required a long and slow migration of data among various servers and temporary tables. A lot of the slowness was just us being super careful, and we were largely able to continue normal operations during most of this time. Anyway I literally just dropped the last temporary table and its database spaces a few hours ago. Check that off!

One of the temporary servers used for the above is now being repurposed as a desperately needed file server just for internal/sysadmin use (temporary storage for database backups, scratch space, etc.). For this I just spent a couple hours last week unloading and reloading slightly bigger hard drives in 48 drive trays.

A couple months ago we also checked another big thing off our list: getting off of Hurricane Electric and going back to using the campus network for our whole operation. The last time campus supported all our bandwidth needs was around 2002 (when the whole campus had a 100Mbit limit, and paid for bandwidth by the bit). The upshot of this is that we no longer have to pay for our own bandwidth ($1000/month) and we can also manage our own address space instead of relying on campus. Basically it's all much cheaper and easier to maintain now. Plus we're also no longer relying on various routers out of our control including one at the PAIX that we haven't been able to log into for years.

But! Of course there were a couple unexpected snags with this network change. First, our lustre file server had a little fit when we changed its IP addresses to this new address space. So we changed it back, but it still wouldn't work! Long story short, we learned a lot about lustre and the voodoo necessary to keep these things behaving. Making matters more confusing was a switch that was part of this lustre complex having its own little fit.

The other snag was moving some campus management addresses into our address space, which also should have been trivial, but unearthed this maddening, and still not completely understood, problem where one of the two routers directing all the traffic in and out of the campus data center seemed unhappy with a small random subset of our addresses, and people all over the planet were intermittently unable to reach our servers. I think the eventual solution was campus rebooting the problem router.

Those starving for new Astropulse work - I swear new data from Arecibo will be coming. Just waiting for enough disks to make a complete shipment. Meanwhile Jeff is hard at work making a Green Bank splitter. Lots of fresh data from fresh sources coming around the bend... Part of the reason I bumped up the ceiling for results-ready-to-send was to do a little advance stress testing on this front.

Oh yeah there was that boinc bug (in the web code) that caused the mysql replica to break every so often. Looks like that's fixed.

Over the weekend lots of random servers had headaches due to one of the GALFA machines going down. It's on the list to separate various dependencies such that this sort of thing doesn't keep happening. Didn't help that me and Eric were both on vacation when this went down.

Meanwhile my daily routine includes a large helping of SERENDIP 6 development, a lesser helping of messing around with VMs (as we start entering the modern age), and taking care of various bits of the Breakthrough Listen project that have fallen on my plate.

- 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: 1740982 · Report as offensive
Profile SciManStev Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 20 Jun 99
Posts: 6651
Credit: 121,090,076
RAC: 0
United States
Message 1740987 - Posted: 10 Nov 2015, 0:51:40 UTC

Matt, that information is more important to us than gold itself.
Thank you!

Steve
Warning, addicted to SETI crunching!
Crunching as a member of GPU Users Group.
GPUUG Website
ID: 1740987 · Report as offensive
Profile mr.mac52
Avatar

Send message
Joined: 18 Mar 03
Posts: 67
Credit: 245,882,461
RAC: 0
United States
Message 1740990 - Posted: 10 Nov 2015, 1:02:37 UTC

Thank you Matt, your reporting is absolutely wonderful for us crunchers outside of Berkeley and greatly appreciated!

John
ID: 1740990 · Report as offensive
Claggy
Volunteer tester

Send message
Joined: 5 Jul 99
Posts: 4654
Credit: 47,537,079
RAC: 4
United Kingdom
Message 1740991 - Posted: 10 Nov 2015, 1:03:27 UTC - in response to Message 1740982.  

Thanks for your update Matt,

Claggy
ID: 1740991 · Report as offensive
Profile Shawn Rothermund
Volunteer tester
Avatar

Send message
Joined: 13 Feb 03
Posts: 132
Credit: 79,997,445
RAC: 123
United States
Message 1741179 - Posted: 10 Nov 2015, 23:58:40 UTC

Thanks so much for the update Matt and am glad to here that there is progress on everything you are working on. Will wait patiently for anything new. Thanks again.
ME AND MY BOY LOOKING FOR ET
ID: 1741179 · Report as offensive
Profile Jimbocous Project Donor
Volunteer tester
Avatar

Send message
Joined: 1 Apr 13
Posts: 1849
Credit: 268,616,081
RAC: 1,349
United States
Message 1742173 - Posted: 15 Nov 2015, 0:29:14 UTC - in response to Message 1740991.  

Thanks for your update Matt,

Claggy

+1 !!
ID: 1742173 · Report as offensive
Profile BQL_FFM
Volunteer tester
Avatar

Send message
Joined: 25 Jan 04
Posts: 11
Credit: 1,197,409
RAC: 0
Germany
Message 1755513 - Posted: 11 Jan 2016, 5:00:43 UTC - in response to Message 1740982.  

Thanks for your update and your work !

Joachim
ID: 1755513 · Report as offensive
raydar115

Send message
Joined: 6 Oct 02
Posts: 17
Credit: 16,305,128
RAC: 0
United States
Message 1757284 - Posted: 18 Jan 2016, 7:56:25 UTC

Thanks Matt,
Any idea of a ETA for cuda to be released for V8 (and i don't mean the tomato juice)
ID: 1757284 · Report as offensive
kittyman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 9 Jul 00
Posts: 51468
Credit: 1,018,363,574
RAC: 1,004
United States
Message 1757285 - Posted: 18 Jan 2016, 7:58:50 UTC - in response to Message 1757284.  

Thanks Matt,
Any idea of a ETA for cuda to be released for V8 (and i don't mean the tomato juice)

Do some reading in the NC forum.....
The subject is being addressed and shall be first announced there when ready.
Coming, hopefully, to a GPU near YOU.
"Freedom is just Chaos, with better lighting." Alan Dean Foster

ID: 1757285 · Report as offensive
raydar115

Send message
Joined: 6 Oct 02
Posts: 17
Credit: 16,305,128
RAC: 0
United States
Message 1757353 - Posted: 18 Jan 2016, 21:29:36 UTC - in response to Message 1757285.  

Thanks
ID: 1757353 · Report as offensive
paul

Send message
Joined: 31 Jan 02
Posts: 4
Credit: 49,832
RAC: 0
United States
Message 1759037 - Posted: 24 Jan 2016, 16:05:51 UTC - in response to Message 1757285.  

Which forum is best for answering this q :

Will Boinc for gpu run on amd radeon gpu's?


If so, will my computer be able to run boinc simultaneously on it's amd CPU
and GPU ?

Thanks for responding..
ID: 1759037 · 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 1759042 - Posted: 24 Jan 2016, 16:25:30 UTC - in response to Message 1759037.  

ID: 1759042 · Report as offensive
Profile tullio
Volunteer tester

Send message
Joined: 9 Apr 04
Posts: 8797
Credit: 2,930,782
RAC: 1
Italy
Message 1759088 - Posted: 24 Jan 2016, 19:37:04 UTC - in response to Message 1759037.  

SETI tasks run my Radeon GPUs (one onboard and one discrete) on my Windows 10 PC. But the error rate is rather high on stock apps. I have installed the Lunatics app and am watching what happens. If one task takes a long time it will result in a computation error.
Tullio
ID: 1759088 · Report as offensive

Message boards : Technical News : Catchup (Nov 09 2015)


 
©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.