Questions and Answers :
GPU applications :
ATI GPU loops at 21.611 % for almost all units
Message board moderation
Author | Message |
---|---|
Govert Schalk Send message Joined: 22 Feb 02 Posts: 1 Credit: 1,624,275 RAC: 1 |
My ATI GPU Seti applications run till 21,611% completed ( about 30 different units), and then end up in an endless loop from 32 minutes till 8 hours and never finish. Is there a bug in the GPU software, or is there a problem with my settings? |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
I am having the same difficulty with my ati gpu. I am running a radon hd 6570, have plenty of power of 730 watts, using a intel i7 cpu. I have also update my catalyst contol to the latest version. I hope someone my respond to our concerns. I was going to add another 6670, but now that doesn't seem reasonable. Thanks for any information. merle - vote yes for freedom of speech |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
There have been reports of problems with the ATI's. Supposedly they were working on new apps that don't use the CAL drivers for the GPUs. Look at the thread about 3 down and read thru it. Since I don't use AMD GPUs I'm not as familiar with them and the issues they are having. However, I do remember them talking about no support from AMD after a certain version of the catalyst driver (i think it was 12.2) One of the work arounds was to use Lunatics optimization programs that did away with the requirements for the catalyst driver. Might be something to look at. The only other thing is to post your first 30 lines of the advance event log for use to review but since it does actually start I don't think that will show us much. Someone else will probably have some more ideas Zalster |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
I just installed the lunatics program for seti but it failed again at the 21.xx% complete point. Anyone else have any ideas? Thanks again for your time and consideration. merle - vote yes for freedom of speech |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
first 30 lines of event log: 6/27/2014 11:27:05 AM | | cc_config.xml not found - using defaults 6/27/2014 11:27:05 AM | | Starting BOINC client version 7.2.42 for windows_x86_64 6/27/2014 11:27:05 AM | | log flags: file_xfer, sched_ops, task 6/27/2014 11:27:05 AM | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6 6/27/2014 11:27:05 AM | | Data directory: C:\ProgramData\BOINC 6/27/2014 11:27:05 AM | | Running under account merle 6/27/2014 11:27:05 AM | | CAL: ATI GPU 0: AMD Radeon HD 6570/6670/7570/7670 series (Turks) (CAL version 1.4.1848, 1024MB, 991MB available, 1248 GFLOPS peak) 6/27/2014 11:27:05 AM | | OpenCL: AMD/ATI GPU 0: AMD Radeon HD 6570/6670/7570/7670 series (Turks) (driver version 1445.5 (VM), device version OpenCL 1.2 AMD-APP (1445.5), 1024MB, 991MB available, 1248 GFLOPS peak) 6/27/2014 11:27:05 AM | | OpenCL CPU: Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz (OpenCL driver vendor: Advanced Micro Devices, Inc., driver version 1445.5 (sse2), device version OpenCL 1.2 AMD-APP (1445.5)) 6/27/2014 11:27:05 AM | SETI@home | Found app_info.xml; using anonymous platform 6/27/2014 11:27:05 AM | | Host name: 81vo 6/27/2014 11:27:05 AM | | Processor: 8 GenuineIntel Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz [Family 6 Model 26 Stepping 4] 6/27/2014 11:27:05 AM | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 cx16 sse4_1 sse4_2 popcnt syscall nx lm vmx tm2 pbe 6/27/2014 11:27:05 AM | | OS: Microsoft Windows 8.1: Core x64 Edition, (06.03.9600.00) 6/27/2014 11:27:05 AM | | Memory: 3.00 GB physical, 6.00 GB virtual 6/27/2014 11:27:05 AM | | Disk: 224.27 GB total, 190.47 GB free 6/27/2014 11:27:05 AM | | Local time is UTC -4 hours 6/27/2014 11:27:05 AM | Collatz Conjecture | URL http://boinc.thesonntags.com/collatz/; Computer ID 136638; resource share 1 6/27/2014 11:27:05 AM | SETI@home | URL http://setiathome.berkeley.edu/; Computer ID 7304674; resource share 1 6/27/2014 11:27:05 AM | | General prefs: from http://einstein.phys.uwm.edu/ (last modified 19-Jul-2013 08:20:57) 6/27/2014 11:27:05 AM | | Computer location: home 6/27/2014 11:27:05 AM | | General prefs: no separate prefs for home; using your defaults 6/27/2014 11:27:05 AM | | Reading preferences override file 6/27/2014 11:27:05 AM | | Preferences: 6/27/2014 11:27:05 AM | | max memory usage when active: 2456.07MB 6/27/2014 11:27:05 AM | | max memory usage when idle: 2763.07MB 6/27/2014 11:27:05 AM | | max disk usage: 44.85GB 6/27/2014 11:27:05 AM | | (to change preferences, visit a project web site or select Preferences in the Manager) 6/27/2014 11:27:05 AM | | Not using a proxy 6/27/2014 11:27:07 AM | SETI@home | Sending scheduler request: To fetch work. |
Jord Send message Joined: 9 Jun 99 Posts: 15184 Credit: 4,362,181 RAC: 3 |
Aborting all the work isn't going to help much in this. That doesn't leave any message other than that you have aborted the work in the database. Option: when you see a task hanging at 21%, keep BOINC running, but navigate to the BOINC data directory (C:\ProgramData\BOINC in your case), open the \slots directory, open the numbers directory one by one until you come across the task that's running and hanging, copy the stderr.txt file to elsewhere and open it. Then copy out and post all the contents of that here in the forums. Why not just open it in the directory? Well, if you can, you can do so, but last time I tried to do so, Notepad and Wordpad would tell me that the file was in use (by the Seti application, of course) and that I therefore couldn't open it. |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
This time it stopped running at 20.733%. the stderr.txt is: Running on device number: 0 Priority of worker thread raised successfully Priority of process adjusted successfully, below normal priority class used OpenCL platform detected: Advanced Micro Devices, Inc. BOINC assigns device 0 Info: BOINC provided device ID used Build features: SETI7 Non-graphics OpenCL USE_OPENCL_HD5xxx OCL_ZERO_COPY OCL_CHIRP3 FFTW AMD specific USE_SSE x86 CPUID: Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz Cache: L1=64K L2=256K CPU features: FPU TSC PAE CMPXCHG8B APIC SYSENTER MTRR CMOV/CCMP MMX FXSAVE/FXRSTOR SSE SSE2 HT SSE3 SSSE3 SSE4.1 SSE4.2 OpenCL-kernels filename : MultiBeam_Kernels_r1843.cl ar=0.426100 NumCfft=196209 NumGauss= 1109259990 NumPulse= 119353171455 NumTriplet= 16681706520576 Currently allocated 145 MB for GPU buffers In v_BaseLineSmooth: NumDataPoints=1048576, BoxCarLength=8192, NumPointsInChunk=32768 Windows optimized S@H Enhanced application by Alex Kan Version info: SSEx (Intel, Core 2-optimized v8-nographics) V5.13 by Alex Kan SSEx Win32 Build 1843 , Ported by : Raistmer, JDWhale SETI7 update by Raistmer OpenCL version by Raistmer, r1843 AMD HD5 version by Raistmer Number of OpenCL platforms: 1 OpenCL Platform Name: AMD Accelerated Parallel Processing Number of devices: 1 Max compute units: 6 Max work group size: 256 Max clock frequency: 650Mhz Max memory allocation: 536870912 Cache type: None Cache line size: 0 Cache size: 0 Global memory size: 1073741824 Constant buffer size: 65536 Max number of constant args: 8 Local memory type: Scratchpad Local memory size: 32768 Queue properties: Out-of-Order: No Name: Turks Vendor: Advanced Micro Devices, Inc. Driver version: 1445.5 (VM) Version: OpenCL 1.2 AMD-APP (1445.5) Extensions: cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_gl_sharing cl_ext_atomic_counters_32 cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_image2d_from_buffer_read_only cl_khr_spir cl_khr_gl_event Work Unit Info: ............... Credit multiplier is : 2.85 WU true angle range is : 0.426100 Used GPU device parameters are: Number of compute units: 6 Single buffer allocation size: 64MB max WG size: 256 period_iterations_num=20 Spike: peak=24.32541, time=33.55, d_freq=1419991581.99, chirp=16.07, fft_len=128k Spike: peak=24.68043, time=33.55, d_freq=1419991581.99, chirp=16.075, fft_len=128k Spike: peak=25.61394, time=33.55, d_freq=1419991581.98, chirp=16.077, fft_len=128k Spike: peak=24.80085, time=33.55, d_freq=1419991581.99, chirp=16.081, fft_len=128k |
Jord Send message Joined: 9 Jun 99 Posts: 15184 Credit: 4,362,181 RAC: 3 |
Have you ever managed to run work on that GPU? Do you see (m)any driver restarts? Reading the readme's, I think you could try the following: Don't use the HD5xxx application for MB7, but instead use the Plain version. This one is for those cards that have driver incompatibilities, something that yours seems to have. .. Readme's? Yes, there are readme files in the installer, but it doesn't unpack those. If you have a third party unzip or unrar program you can right-click on the Lunatics_0.41*.exe file and then do Open archive->$_OUTDIR->docs to find them. |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
I can run collatz without any problems. That is a problem involving only integer math if that makes any difference. I will now look to see what I can find on that lunatics reference. Thanks for your help. |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
addendum: I found your reference and uninstalled the old lunatics and reinstalled the new lunatics for the older hd series although it wasn't specifically called "plain". I am now running it to see if it works. By my estimates of completion times, the gpu would take nearly as long as one thread of my 8 total threads, which would be something of a disappointment since I'm used to the kind of speed with Collatz Conjecture. Any comment? Thanks again. |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
The older lunatics still doesn't work with my card. It stopped at 21.7x%. I guess it's time to give it up. Thanks though. I learned about that stderr.txt though too. |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
This was the latest stderr.txt, which doesn't mean anything to me. unning on device number: 0 Priority of worker thread raised successfully Priority of process adjusted successfully, below normal priority class used OpenCL platform detected: Advanced Micro Devices, Inc. BOINC assigns device 0 Info: BOINC provided device ID used Build features: SETI7 Non-graphics OpenCL USE_OPENCL_HD5xxx OCL_ZERO_COPY OCL_CHIRP3 FFTW AMD specific USE_SSE x86 CPUID: Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz Cache: L1=64K L2=256K CPU features: FPU TSC PAE CMPXCHG8B APIC SYSENTER MTRR CMOV/CCMP MMX FXSAVE/FXRSTOR SSE SSE2 HT SSE3 SSSE3 SSE4.1 SSE4.2 OpenCL-kernels filename : MultiBeam_Kernels_r1843.cl INFO: can't open binary kernel file: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MultiBeam_Kernels_r1843.clHD5_Turks.bin_V7, continue with recompile... INFO: binary kernel file created WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_524288_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_8_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_16_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_32_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_64_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_128_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_256_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_512_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_1024_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_2048_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_4096_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_8192_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_16384_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_32768_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_65536_r1843.bin, continue with recompile... WARNING: can't open binary kernel file for oclFFT plan: C:\ProgramData\BOINC/projects/setiathome.berkeley.edu\MB_clFFTplan_Turks_131072_r1843.bin, continue with recompile... ar=0.440151 NumCfft=192979 NumGauss= 1073881012 NumPulse= 115655312895 NumTriplet= 16332463210496 Currently allocated 145 MB for GPU buffers In v_BaseLineSmooth: NumDataPoints=1048576, BoxCarLength=8192, NumPointsInChunk=32768 Windows optimized S@H Enhanced application by Alex Kan Version info: SSEx (Intel, Core 2-optimized v8-nographics) V5.13 by Alex Kan SSEx Win32 Build 1843 , Ported by : Raistmer, JDWhale SETI7 update by Raistmer OpenCL version by Raistmer, r1843 AMD HD5 version by Raistmer Number of OpenCL platforms: 1 OpenCL Platform Name: AMD Accelerated Parallel Processing Number of devices: 1 Max compute units: 6 Max work group size: 256 Max clock frequency: 650Mhz Max memory allocation: 536870912 Cache type: None Cache line size: 0 Cache size: 0 Global memory size: 1073741824 Constant buffer size: 65536 Max number of constant args: 8 Local memory type: Scratchpad Local memory size: 32768 Queue properties: Out-of-Order: No Name: Turks Vendor: Advanced Micro Devices, Inc. Driver version: 1445.5 (VM) Version: OpenCL 1.2 AMD-APP (1445.5) Extensions: cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_3d_image_writes cl_khr_byte_addressable_store cl_khr_gl_sharing cl_ext_atomic_counters_32 cl_amd_device_attribute_query cl_amd_vec3 cl_amd_printf cl_amd_media_ops cl_amd_media_ops2 cl_amd_popcnt cl_khr_d3d10_sharing cl_khr_d3d11_sharing cl_khr_dx9_media_sharing cl_amd_image2d_from_buffer_read_only cl_khr_spir cl_khr_gl_event Work Unit Info: ............... Credit multiplier is : 2.85 WU true angle range is : 0.440151 Used GPU device parameters are: Number of compute units: 6 Single buffer allocation size: 64MB max WG size: 256 period_iterations_num=20 Spike: peak=25.27335, time=73.82, d_freq=1421199608.09, chirp=-3.1397, fft_len=128k Spike: peak=25.2642, time=73.82, d_freq=1421199608.09, chirp=-3.1406, fft_len=128k |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
I remember seeing that when I was attempting to get the drivers for my laptop to run. Was installing and removing lunatics but it wouldn't take and was causing error. TBar finally told me this. Did you remember to delete the old binary files so new ones could be built with the new driver? Stop BOINC then in the project folder, setiweb.ssl.berkeley.edu_beta, delete the old files ending in .bin, .bin_V6, and .wisdom. You can also delete the <active_task>...</active_task> section for each erred task not running from the client_state.xml file and the task will go back to 'Ready to start' without a slot folder waiting Ignore the beta ending part. We were over on the beta site and that is where the kernal for the beta was kept. You need to do it in the setiathome folder. They told me that doing the first part allowed the computer to recompile the kernal. I maybe wrong but it looks really close to what I was seeing back then. Something to think about. Anyone else with any ideas? Zalster |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
I cleared out a bin_V7 and a wisdom file. Aborted all the old gpu tasks. Reinstalled lunatics and now I am trying to rerun new gpu tasks, hopeing for the best. Thanks for the idea. Tireless George merle - vote yes for freedom of speech |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
It looks like that worked. My gpu is now running past the old %'s and running about 2.4 times as fast as my old cpu units,but it also looks like my cpu is cranking faster than before with the lunatic pgms in there. Wow, I'm holding my breath. As long as I'm at it I have one more question. If I pop in another 6570 or 6670 which I have room for in my asrock mobo; the vendors seem to be saying that if I buy a converter plug from dvi to vga, even though I don't plug it into a second monitor, that will allow me to use two video cards. Can anyone here confirm that or is there another way? Thanks all you guys for your prior help. |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
Well I run 3-4 GPUs in my computers and only one is plugged into a monitor. The others don't need any kind of connection to work. I believe the only time you need to have a dummy plug is if you have a chip with a built in gpu and there is a written program for it. So short, you don't need anything. However....Adding a second GPU after you already have installed the drivers, might require you to reinstall the lunatics. Can't be sure. I know whenever I either move or add a gpu that I end up having to reinstall the drivers for it to work properly. So you might want to wait a bit before trying anything more. But of course it's up to you. Good luck Zalster |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
thanks for the info. I didn't realize that lunatics was actually considered a driver. In any event that will save me $13 on the dummy plug. I never imagined that my cpu could speed up so much with lunatics. When seti wants to install a new version on my computer I guess I need to pass on that and just look to the lunatics site for anything new from them. Is that a correct assumption. And thanks again. |
Zalster Send message Joined: 27 May 99 Posts: 5517 Credit: 528,817,460 RAC: 242 |
A lot of the guys that designed the optimization programs for lunatics also contribute to the seti programs. So when Seti releases new updates, they are actually coming from the same people and tend to be apps we are already running. The lunatics just tend to be more update versions that later get incorporated into the standard Seti apps. If the lunatic group feels we need to upgrade to any of the programs they will post something. Edit Actually Lunatics isn't a driver. It's an app that doesn't require that driver. I don't know how adding a second GPU will affect your computer, if it will be recognized or not. I run NVIDIAs so I have to manually reinstall the drivers for them but you use AMD so I don't know if you would need to reinstall or not. But figuring that you just got it working, why chance it right now. Give it a couple of days and then you can try if you want. Just my 2 cents. Happy Crunching... Zalster |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
I put in my 2nd 6570 and it works right off. Didn't even have to concern myself with any software/driver updates. Didn't need a dummy plug. My 6570's are XFX products. Thanks again to everybody. This is great but now I'll ask another question. I have another pcie x4 slot available. Is it fast enough to handle the seti gpu work units at the same output as my X16's or how much slow down do you think I would encounter? I'm not sure if they even sell fast enough x4 gpu cards. |
merle van osdol Send message Joined: 23 Oct 02 Posts: 809 Credit: 1,980,117 RAC: 0 |
I forgot to say that my pcie x4 is version 2.0. Thanks |
©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.