Posts by Mikie Tim T

WARNING: This website is obsolete! Please follow this link to get to the new Albert@Home website!

1) Message boards : Problems and Bug Reports : Error -5 and 2021 on all GPU workunits (Message 111851) Posted 11 Feb 2012 by Mikie Tim T Post: I tried rebooting as well with all of the projects of this machine paused. I then resumed Albert, so nothing else had a chance to take up memory beforehand. I got the same results, so apparently that wasn't it. Is there anything else that I could possibly try to get it rolling?
2) Message boards : Problems and Bug Reports : Error -5 and 2021 on all GPU workunits (Message 111835) Posted 7 Feb 2012 by Mikie Tim T Post: I tried running a 1.21 workunit as well to see if anything was different, but to no avail. http://albert.phys.uwm.edu/result.php?resultid=115623
3) Message boards : Problems and Bug Reports : Error -5 and 2021 on all GPU workunits (Message 111833) Posted 7 Feb 2012 by Mikie Tim T Post: OK, but that doesn't seem to be my problem according to the BOINC Messages tab: 2/3/2012 10:33:22 PM\|\|ATI GPU 0: ATI Radeon HD 4600 series (R730) (CAL version 1.4.1664, 1024MB, 992MB available, 1024 GFLOPS peak) 2/3/2012 10:33:22 PM\|\|OpenCL: ATI GPU 0: ATI RV730 (driver version CAL 1.4.1664, device version OpenCL 1.0 AMD-APP (851.4), 512MB, 992MB available) 2/3/2012 10:33:22 PM\|\|ATI GPU 0 is OpenCL-capable Are there any other reasons that this message could come about?
4) Message boards : Problems and Bug Reports : Error -5 and 2021 on all GPU workunits (Message 111806) Posted 4 Feb 2012 by Mikie Tim T Post: Upgraded BOINC to 7.0.14 and Catalyst 12.1, but get a similar error: http://albert.phys.uwm.edu/result.php?resultid=114508
5) Message boards : Problems and Bug Reports : Error -5 and 2021 on all GPU workunits (Message 111773) Posted 28 Jan 2012 by Mikie Tim T Post: Well, 12.1 just came out, so I can save a System Restore point and attempt to upgrade from 11.12 if that would help resolve any issues.
6) Message boards : Problems and Bug Reports : Error -5 and 2021 on all GPU workunits (Message 111735) Posted 23 Jan 2012 by Mikie Tim T Post: Oliver, I have cleaned off all drivers, SDK, etc. from my machine and cleanly installed 11.12 drivers from AMD's website. I also updated to 7.0.8 as the Notices recommended. I suspended the other GPU project, exited and restarted the BOINC client, and took screenshots with GPU-Z as you recommended. At no point did the memory exceed even 100MB, but just terminated exactly after 1 minute of processing. Below are the screenshots as well as the pertinent error logging: Stderr output <core_client_version>7.0.8</core_client_version> <![CDATA[ <message> The specified transform does not match the bitmap's color space. (0x7e5) - exit code 2021 (0x7e5) </message> <stderr_txt> Activated exception handling... [21:37:50][4244][INFO ] Starting data processing... [21:37:50][4244][INFO ] Using OpenCL platform provided by: Advanced Micro Devices, Inc. [21:37:50][4244][INFO ] Using OpenCL device "ATI RV730" by: Advanced Micro Devices, Inc. [21:37:51][4244][WARN ] Kernel "kernelTimeSeriesModulation" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 128 work items [21:37:51][4244][WARN ] Kernel "kernelTimeSeriesResampling" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 128 work items [21:37:51][4244][WARN ] Kernel "kernelTimeSeriesMeanReduction" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 32 work items [21:37:51][4244][WARN ] Kernel "kernelTimeSeriesPadding" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 128 work items [21:37:51][4244][WARN ] Kernel "kernelPowerSpectrum" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 32 work items [21:37:51][4244][WARN ] Kernel "kernelHarmonicSumming" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 32 work items [21:37:51][4244][WARN ] Kernel "kernelFillFloatBuffer" exceeds device-specific maximum work group size (requested: 256)! ------> Reducing kernel's work group size to allowed maximum of: 128 work items [21:37:52][4244][INFO ] Checkpoint file unavailable: status.cpt (No such file or directory). ------> Starting from scratch... [21:37:52][4244][INFO ] Header contents: ------> Original WAPP file: ./p2030.20111110.G39.19-00.79.N.b0s0g0.00100_DM417.60 ------> Sample time in microseconds: 65.4762 ------> Observation time in seconds: 274.62705 ------> Time stamp (MJD): 55875.841895339749 ------> 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): 190544.269901 ------> DEC (J2000): 51556.4934006 ------> Galactic l: 0 ------> Galactic b: 0 ------> Name: G39.19-00.79.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: 417.6 cm^-3 pc ------> Scale factor: 0.105215 [21:38:03][4244][INFO ] Seed for random number generator is -1007215342. [21:38:42][4244][INFO ] Derived global search parameters: ------> f_A probability = 0.08 ------> single bin prob(P_noise > P_thr) = 9.93986e-009 ------> thr1 = 18.4267 ------> thr2 = 21.5421 ------> thr4 = 26.5915 ------> thr8 = 35.0049 ------> thr16 = 49.3672 [21:38:43][4244][ERROR] Error during OpenCL FFT setup (error: -5) [21:38:43][4244][ERROR] Demodulation failed (error: 2021)! 21:38:43 (4244): called boinc_finish </stderr_txt> ]]>
7) Message boards : Problems and Bug Reports : Error -5 and 2021 on all GPU workunits (Message 111438) Posted 29 Nov 2011 by Mikie Tim T Post: I am getting the following on all of the GPU workunits downloaded so far: [00:58:42][4536][ERROR] Error during OpenCL FFT setup (error: -5) [00:58:42][4536][ERROR] Demodulation failed (error: 2021)! Sorry, I haven't read the whole thread yet, just in case someone answered already: this error simply means that your GPU doesn't have enough memory available. You need at least a 512 MB graphics card. In any case, if you see this error, you can't run our current app. Sorry, Oliver Actually, this is a 1GB card, so it greatly exceeds the 512MB requirement for the application. Is there any other reason that this error could occur?
8) Message boards : Problems and Bug Reports : Error -5 and 2021 on all GPU workunits (Message 111289) Posted 17 Nov 2011 by Mikie Tim T Post: One other thing I've noticed in the log at the time the error occurs: 11/17/2011 3:19:43 PM \| Albert@Home \| Computation for task p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3 finished 11/17/2011 3:19:43 PM \| Albert@Home \| Output file p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3_0 for task p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3 absent 11/17/2011 3:19:43 PM \| Albert@Home \| Output file p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3_1 for task p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3 absent 11/17/2011 3:19:43 PM \| Albert@Home \| Output file p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3_2 for task p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3 absent 11/17/2011 3:19:43 PM \| Albert@Home \| Output file p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3_3 for task p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3 absent 11/17/2011 3:19:43 PM \| Albert@Home \| Output file p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3_4 for task p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3 absent 11/17/2011 3:19:43 PM \| Albert@Home \| Output file p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3_5 for task p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3 absent 11/17/2011 3:19:43 PM \| Albert@Home \| Output file p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3_6 for task p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3 absent 11/17/2011 3:19:43 PM \| Albert@Home \| Output file p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3_7 for task p2030.20100913.G48.73+01.03.S.b4s0g0.00000_336_3 absent Does this sound like a directory permissions issue or something? I don't have any issues with the currently running CPU workunit or any other project on this machine.
9) Message boards : Problems and Bug Reports : Error -5 and 2021 on all GPU workunits (Message 111288) Posted 17 Nov 2011 by Mikie Tim T Post: Apparently, the driver update also updated the SDK to 2.5, according to the BOINC message log: 11/17/2011 3:05:49 PM \| \| Processor: 1 AuthenticAMD AMD Athlon(tm) 64 Processor 3200+ [Family 15 Model 44 Stepping 2] 11/17/2011 3:05:49 PM \| \| Processor: 512.00 KB cache 11/17/2011 3:05:49 PM \| \| Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 pni syscall nx lm 3dnowext 3dnow 11/17/2011 3:05:49 PM \| \| OS: Microsoft Windows 7: Ultimate x64 Edition, Service Pack 1, (06.01.7601.00) 11/17/2011 3:05:49 PM \| \| Memory: 2.00 GB physical, 4.00 GB virtual 11/17/2011 3:05:49 PM \| \| Disk: 74.50 GB total, 16.91 GB free 11/17/2011 3:05:49 PM \| \| Local time is UTC -6 hours 11/17/2011 3:05:49 PM \| \| ATI GPU 0: ATI Radeon HD 4600 series (R730) (CAL version 1.4.1607, 1024MB, 992MB available, 960 GFLOPS peak) 11/17/2011 3:05:49 PM \| \| OpenCL: ATI GPU 0: ATI RV730 (driver version CAL 1.4.1607, device version OpenCL 1.0 AMD-APP-SDK-v2.5 (793.1), 1024MB) 11/17/2011 3:05:49 PM \| \| ATI GPU is OpenCL-capable So, apparently it isn't a problem with too old an SDK, anyway. I still have some remnants of the previous 2.3 installation, such as Stream Kernel Analyzer 1.7, Stream Profiler 2.1, and Stream SDK Samples 2.3.0.0. Should I just uninstall these, or go ahead and download and install AMD-APP-SDK-v2.5-Windows-64.exe from AMD's website to overwrite them with the current versions?
10) Message boards : Problems and Bug Reports : Error -5 and 2021 on all GPU workunits (Message 111287) Posted 17 Nov 2011 by Mikie Tim T Post: Well, I don't really want to backlevel everything to 11.9 if OpenCL doesn't work on AGP cards at all. This is currently my only ATI card, so I thought I'd try to help work some kinks out of the new app since it will really open up the throughput of the main project once it gets going along. Does anyone else know of any issues running OpenCL on AGP? Has anyone tried on a reasonably current driver? I still have StreamSDK 2.3 installed from when I was crunching Collatz. Would that have any impact on this issue?
11) Message boards : Problems and Bug Reports : Error -5 and 2021 on all GPU workunits (Message 111279) Posted 17 Nov 2011 by Mikie Tim T Post: I am getting the following on all of the GPU workunits downloaded so far: [00:58:42][4536][ERROR] Error during OpenCL FFT setup (error: -5) [00:58:42][4536][ERROR] Demodulation failed (error: 2021)! This is on http://albert.phys.uwm.edu/results.php?hostid=1369. I just downloaded and installed the 11.11 AGP Hotfix from AMD's website and rebooted a couple of hours ago. I also installed all available Windows Updates as well as 6.13.12. I have disabled AMD Overdrive in case the overclock on the GPU had any affect. It did not. I've disallowed any new work to prevent wasting any more workunits until I can get to the bottom of the issue. Any help would be greatly appreciated.

This material is based upon work supported by the National Science Foundation (NSF) under Grant PHY-0555655 and by the Max Planck Gesellschaft (MPG). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the investigators and do not necessarily reflect the views of the NSF or the MPG.

Copyright © 2024 Bruce Allen for the LIGO Scientific Collaboration