WARNING: This website is obsolete! Please follow this link to get to the new Albert@Home website!
[New release] BRP app v1.23/1.24 (OpenCL) feedback thread |
Message boards :
Problems and Bug Reports :
[New release] BRP app v1.23/1.24 (OpenCL) feedback thread
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · Next
Author | Message |
---|---|
tullio Send message Joined: 22 Jan 05 Posts: 796 Credit: 137,342 RAC: 0 |
Albert@home runs well on my Linux box, all results are validated. I have no GPU.I got some validation error on Einstein@home, on a Gamma-ray pulsar search unit. Tullio |
EselTreiber Send message Joined: 29 Apr 08 Posts: 2 Credit: 48,003 RAC: 0 |
Feedback from Ubuntu 12.04_amd64 with Catalyst 12.4 /HD6950@6870: Boinc: last SVN version. Runs fine, no computation errors if all dependencies are installed. (32bit libraries) 2 Tasks on one GPU give me 90-94% GPU-utilisation with CPU load of 12-14% (Core i7 4.3GHz) per Workunit. Performance is (compared to nvidia) 1/2 of a GTX 470. |
steffen_moeller Send message Joined: 9 Feb 05 Posts: 13 Credit: 397,892 RAC: 0 |
During running AaH the desktop was very sticky, most time I had to wait some seconds before any activity could be performed. This was also during the phases of waiting of the AaH task. The desktop was no longer sticky when the AaH project was suspended. This is a very uncomfortable way of operation. ... uncomfortable, but caused by the graphics card interfering with your regular display and is not a defect by albert@home from what I grasp. I observe this with my graphics card on Linux, too. The only way out that I am aware of is to not allow GPU computing while the machine is in use. How much RAM does your card have, btw? I do not observe this behaviour on a 1GB ATI HD 5670 card running albert on Windows, but I do with a HD 5770 512MB card (running prime grid or so because of memory constrains) and this is very much unbearable. Anyone dual booting and observing the issue under Linux but not with Windows? Steffen |
Christoph Send message Joined: 25 Aug 05 Posts: 48 Credit: 208,211 RAC: 0 |
Hi, I have two more errornous wu: http://albert.phys.uwm.edu/result.php?resultid=201372 and http://albert.phys.uwm.edu/result.php?resultid=201360 They have both the same exit code: [23:54:11][5900][ERROR] Error during OpenCL kernel setup: PS_R3 (error: -55) [23:54:11][5900][ERROR] Demodulation failed (error: 2019)! It is a bit different from my last failure. I just told BM to copy all Messages in case you need more info. Hope it works, atm BM is hanging and using one full core and around 700mb memory........ EDIT: Looks like I need to kill BOINC. Still stuck. The export did not happen. Which was that file where the messages are safed? EDIT 2: So it was 'only the Manager that crashed. When I start BoincTask it told me that 4 tasks are running.........Somebody know an AddOn which is saving the Messages to a file outside BOINC? Christoph |
astro-marwil Send message Joined: 28 May 05 Posts: 47 Credit: 1,633 RAC: 0 |
Hallo Steffen! Thank you for your response. ... but caused by the graphics card interfering with your regular display and is not a defect by albert@home from what I grasp. This task was running on a GTX550Ti with 1 GB of RAM in slot 0. At the same time a task of BRP4 from EaH was running on the same card - 0,5 mode -. So you are probably right. I didn´t check for the memory load of the GPU, as in EaH I can easily run 3 task a time. I don´t know, how much of memory the OpenCl task does require. The probably too high memory load might also the reason for the long run time. I will take attention on that next time. Thank you for this hint. Kind regards martin |
Infusioned Send message Joined: 11 Feb 05 Posts: 45 Credit: 149,000 RAC: 0 |
p2030.20110421.G41.18+00.30.N.b6s0g0.00000_1832_2 using einsteinbinary_BRP4 version 123 (atiOpenCL) CPU usage is up a little (steady at ~16% [.16*4cores = ~64%]), but so is GPU usage (45%). All in all, everything is looking good. http://img585.imageshack.us/img585/6087/b6s0g00000018322.jpg |
Infusioned Send message Joined: 11 Feb 05 Posts: 45 Credit: 149,000 RAC: 0 |
p2030.20110421.G41.18+00.30.N.b6s0g0.00000_1400_4 using einsteinbinary_BRP4 version 123 (atiOpenCL) http://img842.imageshack.us/img842/3608/b6s0g00000014004.jpg |
Infusioned Send message Joined: 11 Feb 05 Posts: 45 Credit: 149,000 RAC: 0 |
This wu seems to be wreaking havoc. I completed it ok, but everyone is erroring out. Your client erorred too Bikeman, but I presume that is because you client is 6.12.33? http://albert.phys.uwm.edu/workunit.php?wuid=69493 So far: atiOpenCL: (mine) Completed ok. atiOpenCL: <core_client_version>7.0.27</core_client_version> <![CDATA[ <message> P�i odstra�ov�n� transformace barev do�lo k chyb�. (0x7e3) - exit code 2019 (0x7e3) </message> BRP3Cuda32: <core_client_version>6.12.33</core_client_version> <![CDATA[ <message> - exit code -1073741819 (0xc0000005) </message> atiOpenCL: <core_client_version>7.0.26</core_client_version> <![CDATA[ <message> P�i odstra�ov�n� transformace barev do�lo k chyb�. (0x7e3) - exit code 2019 (0x7e3) </message> <stderr_txt> |
Infusioned Send message Joined: 11 Feb 05 Posts: 45 Credit: 149,000 RAC: 0 |
This wu seems to be wreaking havoc. I completed it ok, but everyone is erroring out. Your client erorred too Bikeman, but I presume that is because you client is 6.12.33? Seems to be the same types of problems with this wu also: http://albert.phys.uwm.edu/workunit.php?wuid=69486 |
ahorek's team Send message Joined: 16 Dec 05 Posts: 2 Credit: 135,508 RAC: 0 |
Got same errors on my notebook with Mobile Radeon 5450 1GB vram: Result: http://albert.phys.uwm.edu/result.php?resultid=204994 I'm using the newest drivers 1.4.1720 and Boinc Client 7.0.27. Previous versions of albert app works. On my another machine with Radeon 5650, there is no problem. Runtime is about 4,5h/wu and memory consumtion 450MB, load 90% with dedicated CPU core (without it only 30%). Log: <core_client_version>7.0.27</core_client_version> <![CDATA[ <message> P�i odstra�ov�n� transformace barev do�lo k chyb�. (0x7e3) - exit code 2019 (0x7e3) </message> <stderr_txt> Activated exception handling... [13:48:03][3088][INFO ] Starting data processing... [13:48:04][3088][INFO ] Using OpenCL platform provided by: Advanced Micro Devices, Inc. [13:48:04][3088][INFO ] Using OpenCL device "Cedar" by: Advanced Micro Devices, Inc. [13:48:05][3088][WARN ] Kernel "kernelTimeSeriesMeanReduction" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 128 work items [13:48:05][3088][WARN ] Kernel "kernelPowerSpectrum" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 128 work items [13:48:05][3088][WARN ] Kernel "kernelHarmonicSumming" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 128 work items [13:48:06][3088][INFO ] Checkpoint file unavailable: status.cpt (No such file or directory). ------> Starting from scratch... [13:48:06][3088][INFO ] Header contents: ------> Original WAPP file: ./p2030.20110421.G41.18+00.30.N.b6s0g0.00000_DM192.00 ------> Sample time in microseconds: 65.4762 ------> Observation time in seconds: 274.62705 ------> Time stamp (MJD): 55672.41520535187 ------> Number of samples/record: 0 ------> Center freq in MHz: 1214.289551 ------> Channel band in MHz: 0.33605957 ------> Number of channels/record: 960 ------> Nifs: 1 ------> RA (J2000): 190551.040699 ------> DEC (J2000): 73613.7874002 ------> Galactic l: 0 ------> Galactic b: 0 ------> Name: G41.18+00.30.N ------> Lagformat: 0 ------> Sum: 1 ------> Level: 3 ------> AZ at start: 0 ------> ZA at start: 0 ------> AST at start: 0 ------> LST at start: 0 ------> Project ID: -- ------> Observers: -- ------> File size (bytes): 0 ------> Data size (bytes): 0 ------> Number of samples: 4194304 ------> Trial dispersion measure: 192 cm^-3 pc ------> Scale factor: 0.00569057 [13:48:13][3088][INFO ] Seed for random number generator is 1158596523. [13:48:57][3088][INFO ] Derived global search parameters: ------> f_A probability = 0.08 ------> single bin prob(P_noise > P_thr) = 1.32531e-008 ------> thr1 = 18.139 ------> thr2 = 21.241 ------> thr4 = 26.2686 ------> thr8 = 34.6478 ------> thr16 = 48.9581 [13:48:58][3088][ERROR] Error during OpenCL kernel setup: PS_R3 (error: -55) [13:48:58][3088][ERROR] Demodulation failed (error: 2019)! 13:48:58 (3088): called boinc_finish </stderr_txt> ]]> |
X1900AIW Send message Joined: 6 May 12 Posts: 2 Credit: 435,065 RAC: 0 |
Software: Catalst 12.3, BOINC 7.0.26 (x64), Windows 7/64 RAM-Usage: Taskmanager during GPU-process: ~207 MB (max) no visible GPU-Usage (by AMD Overdrive), computing the workunits took just same seconds until fail Each workunit failed, so I stopped processing. Stderr output <core_client_version>7.0.26</core_client_version> <![CDATA[ <message> Beim L�schen der Farbtransformation ist ein Fehler aufgetreten. (0x7e3) - exit code 2019 (0x7e3) </message> <stderr_txt> Activated exception handling... [20:24:32][5108][INFO ] Starting data processing... [20:24:33][5108][INFO ] Using OpenCL platform provided by: Advanced Micro Devices, Inc. [20:24:33][5108][INFO ] Using OpenCL device "Cedar" by: Advanced Micro Devices, Inc. [20:24:34][5108][WARN ] Kernel "kernelTimeSeriesMeanReduction" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 128 work items [20:24:34][5108][WARN ] Kernel "kernelPowerSpectrum" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 128 work items [20:24:34][5108][WARN ] Kernel "kernelHarmonicSumming" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 128 work items [20:24:35][5108][INFO ] Checkpoint file unavailable: status.cpt (No such file or directory). ------> Starting from scratch... [20:24:35][5108][INFO ] Header contents: ------> Original WAPP file: ./p2030.20110421.G41.29-00.40.S.b0s0g0.00000_DM42.40 ------> Sample time in microseconds: 65.4762 ------> Observation time in seconds: 274.62705 ------> Time stamp (MJD): 55672.400301627786 ------> Number of samples/record: 0 ------> Center freq in MHz: 1214.289551 ------> Channel band in MHz: 0.33605957 ------> Number of channels/record: 960 ------> Nifs: 1 ------> RA (J2000): 190804.6872 ------> DEC (J2000): 71149.1882019 ------> Galactic l: 0 ------> Galactic b: 0 ------> Name: G41.29-00.40.S ------> Lagformat: 0 ------> Sum: 1 ------> Level: 3 ------> AZ at start: 0 ------> ZA at start: 0 ------> AST at start: 0 ------> LST at start: 0 ------> Project ID: -- ------> Observers: -- ------> File size (bytes): 0 ------> Data size (bytes): 0 ------> Number of samples: 4194304 ------> Trial dispersion measure: 42.4 cm^-3 pc ------> Scale factor: 0.00758342 [20:24:40][5108][INFO ] Seed for random number generator is 1157054464. [20:25:10][5108][INFO ] Derived global search parameters: ------> f_A probability = 0.08 ------> single bin prob(P_noise > P_thr) = 1.32531e-008 ------> thr1 = 18.139 ------> thr2 = 21.241 ------> thr4 = 26.2686 ------> thr8 = 34.6478 ------> thr16 = 48.9581 [20:25:10][5108][ERROR] Error during OpenCL kernel setup: PS_R3 (error: -55) [20:25:10][5108][ERROR] Demodulation failed (error: 2019)! 20:25:10 (5108): called boinc_finish </stderr_txt> ]]> |
Alex Send message Joined: 1 Mar 05 Posts: 88 Credit: 398,734 RAC: 0 |
I gave it a new chance (some weeks ago my system crashed every 20 min). Looks good so far! GPU usage is perfect when running 2 apps at a time CPU usage needs some rework BM 7.0.27 CCC 12.4 edit: figures from the other GPU HD6950 |
Bikeman (Heinz-Bernd Eggenstein) Volunteer moderator Project administrator Project developer Send message Joined: 28 Aug 06 Posts: 1483 Credit: 1,864,017 RAC: 0 |
Hi all Thanks for the testing, we really appreciate it! Some progress report: Today we identified the mysterious cause for the CUDA Windows App 1.24 crashing. We also found and hopefully fixed the problem with some OpenCL app errors (the one with "kernel setup: PS_R3" in the logs). If all goes well the fixed versions will be launched tomorrow, Tuesday, on Albert. All in all we are still "GO" for an OpenCL launch in this or next week :-). Stay tuned. Cheers HB |
Christoph Send message Joined: 25 Aug 05 Posts: 48 Credit: 208,211 RAC: 0 |
This sounds very good! Christoph |
Alex Send message Joined: 1 Mar 05 Posts: 88 Credit: 398,734 RAC: 0 |
Good news! 'I' crunched 7 ATI wu's today, 3 already validated, 4 pending. HD6950: 2 wu's in 1:35 2 GB Ram HD5850: 2 wu's in 2:50 , 1 wu in 1:40 1 GB Ram win7 x 64, i7 2800, 8GB Ram, CCC 12.4, BM 7.0.27 |
ahorek's team Send message Joined: 16 Dec 05 Posts: 2 Credit: 135,508 RAC: 0 |
|
Bikeman (Heinz-Bernd Eggenstein) Volunteer moderator Project administrator Project developer Send message Joined: 28 Aug 06 Posts: 1483 Credit: 1,864,017 RAC: 0 |
looks good, thanks! I don't fully understand the difference in memory usage, but it could be caused by the different capabilities of the cards. Anyone else here with a 54xx ? Cheers HB |
Christoph Send message Joined: 25 Aug 05 Posts: 48 Credit: 208,211 RAC: 0 |
I have a 5450 but not yet the new app. SETI is right now on the GPU. There were some Ghosts wu lingering in my account so I allowed work to get them going. Sometime tomorrow maybe I will pickup new work here. Christoph |
TRuEQ & TuVaLu Send message Joined: 11 Sep 06 Posts: 75 Credit: 615,315 RAC: 0 |
ATI 4850(512MB) no tasks. ATI 5850(1024) running with 0.94cpu and 0.5gpu alongside a milkyway task. progress of task is 44% and ticking. |
Bikeman (Heinz-Bernd Eggenstein) Volunteer moderator Project administrator Project developer Send message Joined: 28 Aug 06 Posts: 1483 Credit: 1,864,017 RAC: 0 |
ATI 4850(512MB) no tasks. Hi! Only OpenCL 1.1 capable cards are supported by this app, that's why the 4850 won't get jobs Cheers HB |