SETI@home v7 v7.03 (opencl_ati_cat132) and Catalyst 13.4


log in

Advanced search

Message boards : Number crunching : SETI@home v7 v7.03 (opencl_ati_cat132) and Catalyst 13.4

Author Message
Stick
Volunteer tester
Send message
Joined: 26 Feb 00
Posts: 84
Credit: 1,645,016
RAC: 566
United States
Message 1405024 - Posted: 19 Aug 2013, 13:36:00 UTC

I recently upgraded from Catalyst 13.1 to 13.4 on my laptop in an attempt to fix a crashing problem with a new app at Collatz. (It worked.) After the upgrade, I started getting SETI@home v7 v7.03 (opencl_ati_cat132) tasks here. Most of the time, these tasks have finished up OK and, as their wingman results come in, they validate. But I've had two results that did not. One crashed at around the 90% complete mark and the other finished up OK but was invalid. My laptop is pretty reliable and is not prone to problems like this. Therefore, I suspect a glitch in software. Any thoughts?
____________

Profile BilBg
Volunteer tester
Avatar
Send message
Joined: 27 May 07
Posts: 2568
Credit: 5,860,006
RAC: 2,221
Bulgaria
Message 1405036 - Posted: 19 Aug 2013, 14:10:22 UTC - in response to Message 1405024.


Do you monitor temperatures of CPU and GPU?
(the 'One' that crashed may be caused by high temperature)

The 'marked as invalid' may be server glitch as the other task of the WU is still 'validation inconclusive':
http://setiathome.berkeley.edu/workunit.php?wuid=1302356833


____________



- ALF - "Find out what you don't do well ..... then don't do it!" :)

Stick
Volunteer tester
Send message
Joined: 26 Feb 00
Posts: 84
Credit: 1,645,016
RAC: 566
United States
Message 1405043 - Posted: 19 Aug 2013, 15:15:32 UTC - in response to Message 1405036.

Do you monitor temperatures of CPU and GPU?
(the 'One' that crashed may be caused by high temperature)

No, I don't monitor temps. But I don't overclock either. And, while it's certainly possible that this one was caused by a hardware hiccup, I'm just not seeing any other indications that point to hardware..

The 'marked as invalid' may be server glitch as the other task of the WU is still 'validation inconclusive':
http://setiathome.berkeley.edu/workunit.php?wuid=1302356833

Hmm! I saw that too, but didn't consider the server glitch possibility. Is that an issue here? If so, I've missed it.
____________

Profile Donald L. Johnson
Avatar
Send message
Joined: 5 Aug 02
Posts: 5994
Credit: 629,608
RAC: 956
United States
Message 1405055 - Posted: 19 Aug 2013, 15:43:39 UTC - in response to Message 1405036.
Last modified: 19 Aug 2013, 15:46:25 UTC


Do you monitor temperatures of CPU and GPU?
(the 'One' that crashed may be caused by high temperature)

The 'marked as invalid' may be server glitch as the other task of the WU is still 'validation inconclusive':
http://setiathome.berkeley.edu/workunit.php?wuid=1302356833

And now it has validated - 2 NVidia GPUs (1 cuda50, 1 Anonymous Platform) got -9 overflow (28,0,0,0,2). Yours must have been different, but the Task Details doesn't give the signal count for yours. Didn't look to see if the others were consistently throwing -9s. Might also be the result file got damaged in transit, that can cause an automatic invalid...

May be just "one of those things". Worry if it happens on a regular basis.
____________
Donald
Infernal Optimist / Submariner, retired

Profile Mike
Volunteer tester
Avatar
Send message
Joined: 17 Feb 01
Posts: 23332
Credit: 31,707,248
RAC: 23,977
Germany
Message 1405076 - Posted: 19 Aug 2013, 16:09:17 UTC

Please delete all files ending with .bin .bin_V7 and .wisdom from your project folder.
The binaries should be created by the new drivers.

____________

Profile BilBg
Volunteer tester
Avatar
Send message
Joined: 27 May 07
Posts: 2568
Credit: 5,860,006
RAC: 2,221
Bulgaria
Message 1405082 - Posted: 19 Aug 2013, 16:17:51 UTC - in response to Message 1405043.


Monitor temperatures:
SIV - System Information Viewer (you need only siv.zip - no install, just unpack and run SIV32X.exe or SIV64X.exe , should work also on your Pentium computers - reads also motherboard sensors, not only CPU internal sensors)
http://rh-software.com/

Control temperatures:
TThrottle (set max temperature for CPU and GPU, BOINC processes will be Throttled if it is reached)
http://efmer.eu/boinc/


____________



- ALF - "Find out what you don't do well ..... then don't do it!" :)

Profile cov_route
Avatar
Send message
Joined: 13 Sep 12
Posts: 286
Credit: 6,112,968
RAC: 15,023
Canada
Message 1405163 - Posted: 19 Aug 2013, 18:47:02 UTC - in response to Message 1405076.

Please delete all files ending with .bin .bin_V7 and .wisdom from your project folder.
The binaries should be created by the new drivers.

+1. Very important and not documented AFAIK except in the forums.

Claggy
Volunteer tester
Send message
Joined: 5 Jul 99
Posts: 4041
Credit: 32,692,280
RAC: 698
United Kingdom
Message 1405173 - Posted: 19 Aug 2013, 19:10:42 UTC - in response to Message 1405163.

Please delete all files ending with .bin .bin_V7 and .wisdom from your project folder.
The binaries should be created by the new drivers.

+1. Very important and not documented AFAIK except in the forums.

In the ReadMe_MultiBeam_OpenCL_ATI.txt Readme:

Known issues:

- Catalyst 12.11 beta and 13.1 have broken OpenCL compiler that will result in driver restarts or invalid results.
But these drivers can be used still if the kernels binaries are precompiled under an older Catalyst driver.
That is, delete all *.bin* files from SETI project directory, revert to Catalyst 12.8 or 12.10, or upgrade to Catalyst 13.2 or later, process at least one task
(check that those *.bin* files were generated again) and (if needed) update to Catalyst 13.1.


Claggy

Profile cov_route
Avatar
Send message
Joined: 13 Sep 12
Posts: 286
Credit: 6,112,968
RAC: 15,023
Canada
Message 1405203 - Posted: 19 Aug 2013, 19:40:01 UTC - in response to Message 1405173.

I'm supposed to read the readme files? Why hasn't anyone explained this to me before!!!

Stick
Volunteer tester
Send message
Joined: 26 Feb 00
Posts: 84
Credit: 1,645,016
RAC: 566
United States
Message 1405208 - Posted: 19 Aug 2013, 19:44:33 UTC - in response to Message 1405163.
Last modified: 19 Aug 2013, 19:48:02 UTC

Please delete all files ending with .bin .bin_V7 and .wisdom from your project folder.
The binaries should be created by the new drivers.

+1. Very important and not documented AFAIK except in the forums.

Mike, Claggy and cov_route,

Thank you for this suggestion! I deleted 47 such files (with varying dates of creation) from my project folder and 20+ have now been recreated. My gut feeling was that the issues I reported had something to do with upgrading Catalyst and at least this suggestion deals with that possibility. Hopefully, the problems will go away and we'll be able to say it worked - but, unfortunately, it's one of those things we won't be able to prove.

Stick
____________

Profile Raistmer
Volunteer developer
Volunteer tester
Avatar
Send message
Joined: 16 Jun 01
Posts: 3368
Credit: 46,067,761
RAC: 30,752
Russia
Message 1405212 - Posted: 19 Aug 2013, 19:50:36 UTC - in response to Message 1405203.

I'm supposed to read the readme files? Why hasn't anyone explained this to me before!!!

LoL, RTFM ;D

____________

Message boards : Number crunching : SETI@home v7 v7.03 (opencl_ati_cat132) and Catalyst 13.4

Copyright © 2014 University of California