BRP4 1.33/1.34 app release: feedback thread

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3,516
Credit: 475,726,558
RAC: 137,296
Topic 196769

Please use this thread to discuss the newly released BRP4 applications v1.33/v1.34.

New features and improvements:
* can process both compressed and uncompressed data files.

Note: v1.33 is technically identical to v1.34 but covers different platforms.

The release of this app version will allow us to switch to compressed input files which has obvious advantages both for the participants of E@H and for the project's infrastructure. If all goes well this will happen next week already.

Cheers,
HB

(retired account)
(retired account)
Joined: 28 Sep 11
Posts: 16
Credit: 7,357,648
RAC: 0

BRP4 1.33/1.34 app release: feedback thread

Just noticed the download of a new BRP4 v1.34 application and thought I should stop by the forum to see what's in it...

This is great news, thanks for the work! Since even my desktops are online with a LTE connection w/ monthly limit, the size of the downloads always limited my participation a bit. Now a GTX 560 Ti, a GT 650M, a HD7950 and a HD7970M should be joining the search for binary pulsars more often!

So far einsteinbinary_BRP4_1.34_windows_x86_64__opencl-ati.exe is running smoothly here, looking forward to the introduction of the compressed data files.

Cheers and a nice weekend to everyone!

Mark my words and remember me. - 11th Hour, Lamb of God

Jeroen
Jeroen
Joined: 25 Nov 05
Posts: 379
Credit: 738,939,197
RAC: 980

This is great news. I have

This is great news.

I have version 1.34 tasks in the queue currently. I am just waiting for the 1.31 tasks to finish up which should be today.

Once compression is enabled, I can setup additional GPUs to run via this project and I may look at purchasing additional AMD hardware since these cards run so well with BRP4.

Thanks.

archae86
archae86
Joined: 6 Dec 05
Posts: 3,036
Credit: 5,263,913,501
RAC: 5,050,269

While I believe these are low

While I believe these are low risk, just to contribute an early look I nudged two 1.33 tasks to run on each of my two GTX460 GPU hosts. They all completed without observable anomaly, with normal execution time and CPU consumption, and already three of the four lucky enough to have fast-responding quorum partners have validated.

I join in celebrating the forthcoming bandwidth reduction. I personally was near half my monthly Comcast data transfer allocation with just two GTX460s. At the moment Comcast has disabled these caps in most areas, and also I have not been able to view my usage for a couple of months now, but I'm glad to have more headroom, just the same.

Neil Newell
Neil Newell
Joined: 20 Nov 12
Posts: 176
Credit: 169,699,457
RAC: 0

Certainly looking good

Certainly looking good here;100+ v1.33 tasks processed (nv270 linux 32-bit) and validated, no errors so far.

Alex
Alex
Joined: 1 Mar 05
Posts: 451
Credit: 406,453,578
RAC: 529,052

I have the first 2 errors

I have the first 2 errors with 1.33 cuda wu*s.
Download failed.
One is http://einsteinathome.org/workunit/146345257

Next one I got was downloaded successful.

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5,516
Credit: 69,080,826,685
RAC: 60,110,393

RE: I have the first 2

Quote:

I have the first 2 errors with 1.33 cuda wu*s.
Download failed.
One is http://einsteinathome.org/workunit/146345257

Next one I got was downloaded successful.


This is not a problem with the app or even with the particular tasks that were affected. The actual stderr output is

7.0.38

WU download error: couldn't get input files:

stochastic_full.bank
-200

]]>


and from this you can see that the problem was with a file that is required for all BRP4 tasks in your cache - stochastic_full.bank.

It's not clear why this file should suddenly be missing (or corrupt) as you seem to have other BRP4 tasks in your cache at the time and that would guarantee that BOINC would be keeping that file at all times. It would normally be deleted if you had no BRP4 tasks on board but would be downloaded again if you then got further new BRP4 tasks. If you browse the file stdoutdae.txt in your BOINC data directory, you may find additional clues as to what happened to the file so as to cause BOINC to trash two tasks that depended on it.

Cheers,
Gary.

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5,516
Credit: 69,080,826,685
RAC: 60,110,393

RE: The release of this app

Quote:
The release of this app version will allow us to switch to compressed input files which has obvious advantages both for the participants of E@H and for the project's infrastructure. If all goes well this will happen next week already.


Sorry to bother you but I'm sure a lot of us are quite anxious to know if "all is going well", and if there might be an update on when the compressed data is expected to flow :-).

Cheers,
Gary.

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4,034
Credit: 220,660,092
RAC: 41,489

We issued the new app

We issued the new app versions on Friday to have a few results to look at after the weekend.

It seems that the new app versions are behaviong ok (well, not worse than the previous ones ;-)) and so we started to incorporate compression into the workunit generation.

Tasks from the first three compressed beams are being sent out right now (beams are named p2030.20110112.G194.37-00.87.N.b4s0g0.00000, p2030.20110112.G194.37-00.87.N.b2s0g0.00000 and p2030.20110112.G194.37-00.87.N.b3s0g0.00000). New work that is pre-processed from now on will also be compressed, but there are still ~700 beams in the buffer that aren't.

We put the old-style, uncompressed beams on hold for now to test the compressed ones during this week, before a few people will be on vacation and we can't change much. At some point (possibly next week) we will have to send these out as well.

Compressing the old beams on the download servers would stress these already busy machines too much. Maybe I find the time to add some reasonable throtteling to the scripts so I can still do that, otherwise I'll send these uncompressed. So don't rely on getting only compressed data from now on yet.

BM

Edit: here's another compressed one: p2030.20120224.G192.84-03.16.S.b4s0g0

BM

DF1DX
DF1DX
Joined: 14 Aug 10
Posts: 72
Credit: 2,242,651,607
RAC: 2,292,458

I got a download error on one

I got a download error on one of the compressed workunits, each size about 720k:

BRP4 1.33 on Win7/ Boinc 7.0.28(x64), i5-760+GTX460

28.01.2013 14:10:28 | Einstein@Home | Requesting new tasks for NVIDIA
28.01.2013 14:10:30 | Einstein@Home | Finished download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3120.bin4
28.01.2013 14:10:30 | Einstein@Home | Started download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3121.bin4
28.01.2013 14:10:30 | Einstein@Home | [error] File p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3120.bin4 has wrong size: expected 2098320, got 716007
28.01.2013 14:10:30 | Einstein@Home | [error] Checksum or signature error for p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3120.bin4
28.01.2013 14:10:32 | Einstein@Home | Finished download of p2030.20110112.G194.62-00.41.N.b5s0g0.00000_3719.bin4
28.01.2013 14:10:32 | Einstein@Home | Started download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3122.bin4
28.01.2013 14:10:32 | Einstein@Home | Scheduler request completed: got 1 new tasks
28.01.2013 14:10:33 | Einstein@Home | Finished download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3121.bin4
28.01.2013 14:10:33 | Einstein@Home | Started download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3123.bin4
28.01.2013 14:10:33 | Einstein@Home | [error] File p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3121.bin4 has wrong size: expected 2098320, got 715671
28.01.2013 14:10:33 | Einstein@Home | [error] Checksum or signature error for p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3121.bin4
28.01.2013 14:10:34 | Einstein@Home | Finished download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3122.bin4
28.01.2013 14:10:34 | Einstein@Home | Started download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3124.bin4
28.01.2013 14:10:34 | Einstein@Home | [error] File p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3122.bin4 has wrong size: expected 2098320, got 715934
28.01.2013 14:10:34 | Einstein@Home | [error] Checksum or signature error for p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3122.bin4
28.01.2013 14:10:35 | Einstein@Home | Finished download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3123.bin4
28.01.2013 14:10:35 | Einstein@Home | Started download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3125.bin4
28.01.2013 14:10:35 | Einstein@Home | [error] File p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3123.bin4 has wrong size: expected 2098320, got 716403
28.01.2013 14:10:35 | Einstein@Home | [error] Checksum or signature error for p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3123.bin4
28.01.2013 14:10:37 | Einstein@Home | Finished download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3124.bin4
28.01.2013 14:10:37 | Einstein@Home | Finished download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3125.bin4
28.01.2013 14:10:37 | Einstein@Home | Started download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3126.bin4
28.01.2013 14:10:37 | Einstein@Home | Started download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3127.bin4
28.01.2013 14:10:37 | Einstein@Home | [error] File p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3124.bin4 has wrong size: expected 2098320, got 716290
28.01.2013 14:10:37 | Einstein@Home | [error] Checksum or signature error for p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3124.bin4
28.01.2013 14:10:37 | Einstein@Home | [error] File p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3125.bin4 has wrong size: expected 2098320, got 716345
28.01.2013 14:10:37 | Einstein@Home | [error] Checksum or signature error for p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3125.bin4
28.01.2013 14:10:40 | Einstein@Home | Finished download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3126.bin4
28.01.2013 14:10:40 | Einstein@Home | Started download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000.zap
28.01.2013 14:10:40 | Einstein@Home | [error] File p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3126.bin4 has wrong size: expected 2098320, got 715407
28.01.2013 14:10:40 | Einstein@Home | [error] Checksum or signature error for p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3126.bin4
28.01.2013 14:10:41 | Einstein@Home | Finished download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3127.bin4
28.01.2013 14:10:41 | Einstein@Home | Finished download of p2030.20110112.G194.37-00.87.N.b2s0g0.00000.zap
28.01.2013 14:10:41 | Einstein@Home | Started download of p2030.20111115.G193.84-00.43.C.b3s0g0.00000_3552.bin4
28.01.2013 14:10:41 | Einstein@Home | Started download of p2030.20111115.G193.84-00.43.C.b3s0g0.00000_3553.bin4
28.01.2013 14:10:41 | Einstein@Home | [error] File p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3127.bin4 has wrong size: expected 2098320, got 715945
28.01.2013 14:10:41 | Einstein@Home | [error] Checksum or signature error for p2030.20110112.G194.37-00.87.N.b2s0g0.00000_3127.bin4

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4,034
Credit: 220,660,092
RAC: 41,489

Thanks for the report.

Thanks for the report. Something weird and ugly must have happened to this beam when compressing, I'm trying to fix that.

BM

BM

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.