Posts by [AF>EDLS]GuL

1) Message boards : News : GPU Issues Mega Thread (Message 66090) Posted 7 Jan 2017 by [AF>EDLS]GuL Post: Great ! Happy crunching and happy new year !
2) Message boards : News : GPU Issues Mega Thread (Message 66084) Posted 4 Jan 2017 by [AF>EDLS]GuL Post: Thanks for the tips. The first link, in the fist line, clearly indicates that the article applies to AMD Processing Cores. Since I have a NVIDIA, I don't think that applies to me. You're welcome. The first link applies to AMD CPUs, as you have. Don't forget that part of the job is done on CPU. If opencl is not available there, it may fail. Cheers
3) Message boards : News : GPU Issues Mega Thread (Message 66082) Posted 3 Jan 2017 by [AF>EDLS]GuL Post: Could the manufacturer of the CPU make a difference whether a CL program compiles or not? I am not sure but it may be a possibility. I have read somewhere, maybe on seti, that the difficulty of opencl is that it can work everywhere, but it is not automatic. In particular, you have no amd opencl line when boinc is starting. Try to install opencl amd sdk http://developer.amd.com/tools-and-sdks/opencl-zone/amd-accelerated-parallel-processing-app-sdk/. Some further interesting reading: https://askubuntu.com/questions/821341/how-to-enable-cpu-platform-for-opencl-applications, https://wiki.tiker.net/OpenCLHowTo
4) Message boards : News : GPU Issues Mega Thread (Message 66063) Posted 30 Dec 2016 by [AF>EDLS]GuL Post: If you click on my name (wb8ili) on the left of this message - then under the Badges row there is a row Computers - click on "here" - then you should see a list of computers - column 2 is the names - scroll to "Desktop" and then click on "Tasks" in column 1 - then go to Invalid. That's the way I checked but I can only see your hostids but not their names. You're right, it seems to be a computation error. I have no more ideas, sorry.
5) Message boards : News : GPU Issues Mega Thread (Message 66058) Posted 30 Dec 2016 by [AF>EDLS]GuL Post: 1) The Linux version on the BOINC Berkeley website isn't kept up to date very well since the BOINC team lost their funding. If you use the Ubuntu Software installer to install BOINC, it will install 7.6.33. From what I understand, it has some improvements at recognizing GPU versions. The Ubuntu installer will install boinc in a different set of folders and run it under the user-id of I don't know on ubuntu but on fedora or centos, new versions are sometimes a regression : on centos I reverted to 7.2.42 due to partial display problems ; on fedora it's not detecting idle state
6) Message boards : News : GPU Issues Mega Thread (Message 66056) Posted 30 Dec 2016 by [AF>EDLS]GuL Post: I can't see the names of your computers. Do you have a link to a faulty work unit, please ? About the answered questions, everything seems correct.
7) Message boards : News : GPU Issues Mega Thread (Message 66054) Posted 29 Dec 2016 by [AF>EDLS]GuL Post: The first thing I noticed is that you are running an older version of BOINC. You might try updating to the latest version and see if that makes a difference. The second thing I noticed is that your event log shows that the driver version of your GPU is unknown. Here is my event log with similar lines. @draco_seti I agree with the fact that the non-decteted driver is strange. Try to install the driver from your distribution repository. Can you also enter the following command in milkyway directory : ldd * \| grep -i "not found"
8) Message boards : News : GPU Issues Mega Thread (Message 66051) Posted 29 Dec 2016 by [AF>EDLS]GuL Post: I don't think the combination of LINUX and NVIDIA cards has worked (returned valid results) since the tasks were bundled. At least mine haven't. I don't know if it working on all cards, but on one of mines it's ok : https://milkyway.cs.rpi.edu/milkyway/result.php?resultid=1923858198
9) Message boards : News : GPU Issues Mega Thread (Message 66044) Posted 28 Dec 2016 by [AF>EDLS]GuL Post: If you can, test them under windows with furemark and check their VRM temperatures with gpu-z. I don't know any linux software providing this information. You can also test your cards on a computer you know working on milky.
10) Message boards : News : GPU Issues Mega Thread (Message 66042) Posted 28 Dec 2016 by [AF>EDLS]GuL Post: Are your cards overclocked ? Milky is very hardware demanding. Can you describe their ventilation ? If you can add a lateral intake, it will help to cool down your VRM (Voltage Regulation Module) and memory ram. Edit : for drivers, prefer distribution ones rather than the ones on nvidia site. You will avoid compatibility problems
11) Message boards : News : GPU Issues Mega Thread (Message 66038) Posted 28 Dec 2016 by [AF>EDLS]GuL Post: @draco_seti It may come from your side : I just tried with a quadro 4000 on fedora 24 with driver 367.57 and it worked perfectly. Have you tried to reset the project ?
12) Message boards : News : GPU Issues Mega Thread (Message 65095) Posted 1 Sep 2016 by [AF>EDLS]GuL Post: I forgot : on the R9 380, I needed to add a app_info.xml in order to recognize the card, and a app_config.xml to compute 4 tasks simultaneously. Good night
13) Message boards : News : GPU Issues Mega Thread (Message 65094) Posted 1 Sep 2016 by [AF>EDLS]GuL Post: @Keith Indeed. And moreover recent AMD have less double precision capacity than old ones. I was really surprised to decrease my Milky power by almost 4 when passing from a HD 7950 to a R9 380. All is working fine for me, except the number of available tasks. Cheers
14) Message boards : Number crunching : AMD Radeon R9 Fury X - app_info.xml and apps - optimizations (Message 65013) Posted 9 Aug 2016 by [AF>EDLS]GuL Post: Hi Jojez J. Have a look there https://milkyway.cs.rpi.edu/milkyway/forum_thread.php?id=3987&postid=64975 Cheers
15) Message boards : Number crunching : app_info.xml Help (Message 64982) Posted 4 Aug 2016 by [AF>EDLS]GuL Post: I'm now fetching work at least but they are still MOSTLY validation inconclusive. (107) with only 2 WU's valid. Any suggestions? Thx Just be patient. The other guy needs also to compute your task, and they must agree. If you have no errors, that's great !
16) Message boards : Number crunching : app_info.xml Help (Message 64980) Posted 4 Aug 2016 by [AF>EDLS]GuL Post: Hi Chooka, All the app_info stuff from Dirk is going into a file called app_info.xml. Be careful that by default, windows extensions are not shown. So your file can in fact be app_info.xml.txt, which is not correct. The app_config.xml is another file to create, but it is only optional. The files .exe have not to be executed alone, but they are needed to crunch. All these files need to be placed into your milkyway directory C:\ProgramData\BOINC\projects\milkyway.cs.rpi.edu_milkyway . This is a hidden directory. Once it's done, you can turn off boinc, not only pausing computation, and then start it again. If everything is right, you should see in the event log references to app_info.xml and app_config.xml for milkyway, and crunch will start. What is your GPU ? Happy crunch ;-)
17) Message boards : Number crunching : app_info.xml Help (Message 64976) Posted 4 Aug 2016 by [AF>EDLS]GuL Post: Hi, For information, here is an interesting discussion on primegrid site, where the project scientist resolved the problem with AMD recent graphic cards : https://www.primegrid.com/forum_thread.php?id=6859#97449 I tried a WU there and it worked. That's maybe an idea ?
18) Message boards : Number crunching : app_info.xml Help (Message 64975) Posted 3 Aug 2016 by [AF>EDLS]GuL Post: Thank you so much Dirk ! It worked perfectly for me on a R9 380 running under windows 10 64 bits and boinc 7.6.22. I just added the following app_config.xml file: <app_config> <app> <name>milkyway</name> <gpu_versions> <gpu_usage>0.25</gpu_usage> <cpu_usage>0.05</cpu_usage> </gpu_versions> </app> </app_config> This way, I can separate applications informations (app_info.xml) and attributed ressources (app_config.xml). Four tasks are done at the same time in 1'40"-1'50", avoiding GPU non-working time when a task is completed. Cheers
19) Message boards : News : Nvidia OpenCL updated (Message 46267) Posted 17 Feb 2011 by [AF>EDLS]GuL Post: All my linux hosts seem to be failing all tasks with the new 0.52 OpenCL app. I made a really dumb mistake in the Linux build. Should be fixed now (0.54). Many thanks, it seems to be working now
20) Message boards : News : Nvidia OpenCL updated (Message 46071) Posted 8 Feb 2011 by [AF>EDLS]GuL Post: I have also the same problem : all the wu 0.52 are failing on my linux ubuntu 10.10 host http://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=66288 I am using nvidia 270.18 x64 with a GTX260, boinc 6.10.58 x64 and have reseted the project. Any help, please ?

Next 20