Posts by Darrell

1) Message boards : Number crunching : Problems running wus on AMD R9 380. Suddenly Moo crashes card. (Message 7735) Posted 5 Oct 2017 by Darrell Post: For being one of the smallest GPU apps, I have found that it is the one that drives the graphics card the hardest out of all of the Boinc projects that use the GPU. When it runs, the temp climbs high, the fans speed up, and the power usage goes to the max. So having to increase the power limit in wattman does not surprise me to get it stable. So if you want to test any overclock settings on your GPU, run a Moo task. If it doesn't fail, the GPU will handle any other project or game you wish to play.
2) Message boards : Number crunching : amd 480 and opencl issues (Message 7706) Posted 28 Jul 2017 by Darrell Post: Ok, it turns out I was reading the dates wrong and your last valid task was on Jul 6th and you were changing the "Distributed.net ATI Stream Client Core" setting in your project preferences. No problem doing this, to find out which works best for you. I use the #3 " IL 4-pipe cs-1 (req. Cypress ASIC or later)" setting. From Your last valid task: 19:17:39 (4308): wrapper v1.4 build 16 for AMD/ATI OpenCL starting (BOINC Wrapper v7.5.26011) 19:17:39 (4308): device: OpenCL: AMD/ATI GPU 0: AMD Radeon (TM) R9 380 Series (driver version 2348.4, device version OpenCL 2.0 AMD-APP (2348.4), 4096MB, 4096MB available, 3548 GFLOPS peak) 19:17:39 (4308): checkpoint interval: 0h15m00s00 (task 1075200 GFLOPS, 0h05m03s03 per packet) 19:17:39 (4308): wrapper: running dnetc520-win32-x86-opencl.exe (-ini dnetc.ini -runoffline -multiok=1) - attempt 1/10 dnetc v2.9111-520-CTR-12082118 for OpenCL on Win32 (WindowsNT 6.1). Using email address (distributed.net ID) '<censored> ' [Jul 06 18:17:41 UTC] Automatic processor type detection did not recognize the processor (tag: "Tonga (AMD Radeon (TM) R9 380 Series)") [Jul 06 18:18:06 UTC] RC5-72: using core #0 (CL ANSI 1-pipe). From Mine: 17:24:34 (6188): wrapper v1.4 build 16 for AMD/ATI OpenCL starting (BOINC Wrapper v7.5.26011) 17:24:34 (6188): device: OpenCL: AMD/ATI GPU 0: Radeon (TM) RX 480 Graphics (driver version 2442.8, device version OpenCL 2.0 AMD-APP (2442.8), 8192MB, 8192MB available, 5990 GFLOPS peak) 17:24:34 (6188): checkpoint interval: 0h15m00s00 (task 1120000 GFLOPS, 0h03m06s96 per packet) 17:24:34 (6188): wrapper: running dnetc520-win32-x86-opencl.exe (-ini dnetc.ini -runoffline -multiok=1) - attempt 1/10 dnetc v2.9111-520-CTR-12082118 for OpenCL on Win32 (WindowsNT 6.1). Using email address (distributed.net ID) '<censored> ' [Jul 28 22:24:35 UTC] RC5-72: using core #3 (CL 4-pipe). Differences due to different AMD GPUs and I have upgraded to latest driver version. From one of your error tasks: 18:26:11 (3544): wrapper v1.4 build 16 for AMD/ATI OpenCL starting (BOINC Wrapper v7.5.26011) 18:26:11 (3544): device: OpenCL: AMD/ATI GPU 0: AMD Radeon (TM) R9 380 Series (driver version 2348.3, device version OpenCL 2.0 AMD-APP (2348.3), 4096MB, 4096MB available, 3548 GFLOPS peak) 18:26:11 (3544): checkpoint interval: 0h15m00s00 (task 1075200 GFLOPS, 0h05m03s03 per packet) 18:26:11 (3544): wrapper: running dnetc520-win32-x86-opencl.exe (-ini dnetc.ini -runoffline -multiok=1) - attempt 1/10 dnetc v2.9111-520-CTR-12082118 for OpenCL on Win32 (WindowsNT 6.1). Using email address (distributed.net ID) '<censored> ' [Jul 20 17:26:12 UTC] Unable to initialize OpenCL. [Jul 20 17:26:12 UTC] Break Shutting down... So the question becomes what has broken Opencl on this device for this project. Is this device used with any other projects to run Opencl tasks? Do the tasks run or fail? Have you tried using GPU-Z to check the status of the device? Are you overclocking or undervolting this device? Try opening a cmd window and switch to the windows\system32 directory and run Clinfo program to verify that Opencl is working.
3) Message boards : Number crunching : amd 480 and opencl issues (Message 7703) Posted 22 Jul 2017 by Darrell Post: I looked at your error report and it's the same as mine. Yours: 00:31:27 (2532): device: OpenCL: AMD/ATI GPU 0: Radeon (TM) RX 480 Graphics (driver version 2348.3, device version OpenCL 2.0 AMD-APP (2348.3), 4096MB, 4096MB available, 6018 Mine: 22:35:31 (3036): device: OpenCL: AMD/ATI GPU 0: AMD Radeon (TM) R9 380 Series (driver version 2348.3, device version OpenCL 2.0 AMD-APP (2348.3), 4096MB, 4096MB available, 3548 Only difference I see is : dnetc v2.9111-520-CTR-12082118 for OpenCL on Win32 (WindowsNT 6.1). on mine. dnetc v2.9111-520-CTR-12082118 for OpenCL on Win32 (WindowsNT 6.2). on yours. I'm running Windows Server 2008 R2 (same as Windows 7). GPU is working on everything else but Moo. Was working fine on Moo until just over a week ago. I'm running 17.4.4 as it was suggested as a fix. You say you removed the Intel cpu OpenCL drivers and this made it work? I'm gonna try the latest drivers again in a seemingly futile attempt to get this working again. I'm goin' nuts!! Your last valid task was on 29 Jun 2017, so it has been failing for more than a week. First suggestion - when a problem arises the first thing to do is to drop your cache sizes to zero and get tasks one at a time until the problem is fixed. Next, look in the boinc data directory for a file coproc_info.xml file and see if the opencl information for your gpu is correct. Next try and remember what changes you may have made to your system on the 29th and back it out. Next, use DDU to remove video driver and re-install a clean version.
4) Message boards : Number crunching : AMD RX 480 (Message 7606) Posted 17 Mar 2017 by Darrell Post: This is the app_config to run two tasks on my new rx480: <app_config> <app> <name>dnetc</name> <max_concurrent>2</max_concurrent> <fraction_done_exact/> <gpu_versions> <gpu_usage>0.5</gpu_usage> <cpu_usage>0.05</cpu_usage> </gpu_versions> </app> </app_config> Two tasks complete in 90 -100 seconds of cpu time. Not too bad considering the cpu is just an Athlon II 259 and the pcie slot is 2.0.
5) Message boards : Number crunching : GPU ATI tasks never start (Message 4585) Posted 20 Feb 2013 by Darrell Post: Anytime, please come back and let us know how it works out for you. The app_config.xml does work and the app name does have to be set to dnetc. Did not lose any tasks from changing from app info to app config, however there are two tasks that have gone missing from the que when I first put in the app info xml. Now it's a matter of using it with other projects to fine tune the system.
6) Message boards : Number crunching : GPU ATI tasks never start (Message 4580) Posted 20 Feb 2013 by Darrell Post: Yes, I have tried 11.9 to 12.1, but amd made some sort of change to the cal routines that causes boinc to see no usable gpus. After 12.1 amd dropped the cal routines for XP completely in favor of open-cl. It's just strange that they keep the cal routines for win7. I will give the app config a try later on today, will probably wipe out the five tasks waiting to run, but on the integrated hd3000 in my win7 machine, the tasks are taking about 15-18 hours to complete and to wait for them all to complete will take days. Thanks for the info, Mikey.
7) Message boards : Number crunching : GPU ATI tasks never start (Message 4574) Posted 19 Feb 2013 by Darrell Post: Just joined today and my XP machine with one usable gpu started processing right off the git go. My win 7 machine with two gpus refused to run any until I put in the first app info xml, here is the startup log after putting in the xml: 2/18/2013 5:54:45 PM \| \| log flags: file_xfer, sched_ops, task, coproc_debug, scrsave_debug, task_debug 2/18/2013 5:54:45 PM \| \| Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6 2/18/2013 5:54:45 PM \| \| Data directory: C:\ProgramData\BOINC 2/18/2013 5:54:45 PM \| \| Running under account Darrell 2/18/2013 5:54:45 PM \| \| Processor: 2 AuthenticAMD AMD Athlon(tm) II X2 250 Processor [Family 16 Model 6 Stepping 3] 2/18/2013 5:54:45 PM \| \| Processor: 1.00 MB cache 2/18/2013 5:54:45 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 htt pni cx16 syscall nx lm svm sse4a osvw ibs skinit wdt page1gb rdtscp 3dnowext 3dnow 2/18/2013 5:54:45 PM \| \| OS: Microsoft Windows 7: Professional x64 Edition, Service Pack 1, (06.01.7601.00) 2/18/2013 5:54:45 PM \| \| Memory: 7.75 GB physical, 15.50 GB virtual 2/18/2013 5:54:45 PM \| \| Disk: 931.41 GB total, 844.43 GB free 2/18/2013 5:54:45 PM \| \| Local time is UTC -6 hours 2/18/2013 5:54:45 PM \| \| VirtualBox version: 4.2.4 2/18/2013 5:54:45 PM \| \| CAL: ATI GPU 0: ATI Radeon HD 5800 series (Cypress) (CAL version 1.4.1741, 1024MB, 991MB available, 4176 GFLOPS peak) 2/18/2013 5:54:45 PM \| \| CAL: ATI GPU 1: ATI Radeon HD 2300/2400/3200 (RV610) (CAL version 1.4.1741, 256MB, 224MB available, 56 GFLOPS peak) 2/18/2013 5:54:45 PM \| \| OpenCL: AMD/ATI GPU 0: ATI Radeon HD 5800 series (Cypress) (driver version CAL 1.4.1741 (VM), device version OpenCL 1.2 AMD-APP (938.2), 1024MB, 991MB available, 4176 GFLOPS peak) 2/18/2013 5:54:45 PM \| \| No NVIDIA library found 2/18/2013 5:54:45 PM \| Moo! Wrapper \| Found app_info.xml; using anonymous platform After getting the tasks to run, I have gone back and put back the exclude gpu in the config file to restrict moo's use to gpu #1 as I want to use gpu #0 for those projects that run opencl. Really was hoping to stay away from using any app info files, but is apparently the only way to get your tasks to work on muligpu system.