Posts by Teemu Mannermaa

61) Questions and Answers : Unix/Linux : Not Returning Results Automatically..?? (Message 6575) Posted 9 Jun 2015 by Teemu Mannermaa Post: With moowrap results just sit there "Ready To Report" until I do a "manual" update Reporting any uploaded results is decided solely by the BOINC Client, neither our apps nor our scheduler have anything to say about the matter. How long have you given the BOINC Client to decide it wants to report our tasks and get new ones? Which of your hosts is affected or are they all behaving similarly? If you have long queue of work, it can take a while but it should eventually decide to do it unless there's a bug. Also, any uploads pending or are these task really ready? It's also possible for BOINC Client to postpone reporting if additional tasks are ready soon. Anything relevant in BOINC Client event logs related to our project regarding decisions on doing a scheduler request? You might need to add some debug flags to see more info. If you can, you can try updating to a version in the 7.4 branch as there has been various changes in different BOINC Client version on when and why to make a scheduler request. However, things should work with 7.2.42 too so it's not mandatory to update. -w
62) Message boards : News : New v1.4 apps for nVidia OpenCL and Android (Message 6560) Posted 1 Jun 2015 by Teemu Mannermaa Post: The Linux versions are on the list next and will be deployed soon. Will this include support for 7950 GPUs? Ah, yes, the new OpenCL app should run on those. Linux apps are now in beta so if you want to give them a spin before I release them for all you can opt in for beta work. Thank you for any tests done (or your patience if you just wanna wait)! :) -w
63) Message boards : News : New v1.4 apps for nVidia OpenCL and Android (Message 6559) Posted 1 Jun 2015 by Teemu Mannermaa Post: It has run the OpenCL app a few times, but keeps reverting back. This host's results are the one in question. Looking through the latest results, it's mostly running the OpenCL app now. It can take a while for the BOINC scheduler to start preferring the faster app and it might once in a while test run the other one to see if things have changed. Hopefully things keep going this way for your host. -w
64) Message boards : News : New v1.4 apps for nVidia OpenCL and Android (Message 6551) Posted 27 May 2015 by Teemu Mannermaa Post: 1.04 steals a whole CPU. I now have a paused CPU project because of it. 1.03 did not do this. Can that please be fixed. I know that OpenCL usually is Right, it was reported to me previously that the OpenCL (especially nVidia one) benefits from having one free CPU for it to feed the GPU and thus is specified as requiring one CPU. However, this does change a lot between Windows and Linux as well different driver versions so it's not that easy to find "correct value". I can surely change it, that's not a problem. I'll make it same as AMD OpenCL app (it requires 0.25 CPU) for now to see how things go. The change should be live within an hour and subsequent work sent should have the changed CPU requirement. Thanks for the report! -w
65) Message boards : News : New v1.4 apps for nVidia OpenCL and Android (Message 6549) Posted 24 May 2015 by Teemu Mannermaa Post: More apps based on v1.4 of wrapper code are now available. First there are Android apps with both non-PIE (for pre-4.1) and PIE (for 4.1 and later, especially 5) versions. Android apps are no longer in beta so everybody should get them but please report any problems at the new Android forum http://moowrap.net/forum_forum.php?id=27. Secondly there's a new nVidia OpenCL app available for Windows. This brings the same level of support for nVidia GPUs that was previously available only for AMD GPUs. The Linux versions are on the list next and will be deployed soon. Happy crunching!
66) Message boards : Wish list : Android (Message 6547) Posted 23 May 2015 by Teemu Mannermaa Post: With BETA apps in the mix, these forums should have an Android section just as it does Windows, Linux, and Mac. Unless of course you feel that Android belongs in the Linux forum... Good idea, thanks for suggesting this! Now there's an Android section for any Android specific problems. -w
67) Message boards : Number crunching : No stats (Message 6543) Posted 22 May 2015 by Teemu Mannermaa Post: No stats are shown there : http://www.allprojectstats.com/showuser.php?projekt=0&id=2633602 Ah, interesting, looks like they have failed to process our exported stats ince August 2014. After looking into this, it seems, unlike all the other stats sites, they want to correctly fetch the db_dump.xml file which was not happening due to permission issues before processing our stats further. That issue is now fixed and according to http://allprojectstats.com/showlog.php?projekt=103 they have managed to process our stats once again. Apologies taking this long to look into this issue, it's unacceptable. :( -w
68) Message boards : Number crunching : NSA backdoors accusation (Message 6542) Posted 21 May 2015 by Teemu Mannermaa Post: If you are still participating in the RC5-72 effort, please stop. It has become counterproductive for a number of reasons. First of all, besides the fact that the challenge has officially been dropped by RSA Labs themselves almost 7(!) years ago now, there is also the inescapable zeitgeist:... While RSA Labs has dropped support for the RC5 challenge Distributed.net has decided to "privately sponsor" the prize and continue running the project. You can read more about this at http://www.distributed.net/RC5. ...to secretly include a backdoor in their products as early as 2004 and therefore taints the company and any ongoing promotion through.. While I won't comment on these wild accusations I can say that there's no RSA Labs code in the upstream apps so there's can't be any RSA backdoor in them. If in doubt, you can audit the public source code available at http://www.distributed.net/source (including a github repo). All in all, I don't see these statements as any reasons to stop contributing to the Distributed.net RC5-72 challenge. -w
69) Questions and Answers : Web site : Login page uses invalid certificate (Message 6215) Posted 4 Oct 2014 by Teemu Mannermaa Post: Browser at home complains, but lets me load the login page, but my work machine's policy prevents logging in at all. This is an old post but as a note, I've renewed our certificate a while back and recent BOINC Server update also fixed any "mixed content" issues. So login page should be fully functional again over SSL. In fact, the whole site should work over SSL now even though BOINC Server code only forces SSL for login. -w
70) Message boards : Number crunching : Beta Android app : validation inconclusive (Message 6212) Posted 29 Sep 2014 by Teemu Mannermaa Post: My Android device (Galaxy Nexus) has just completed its firts WU, but it is in validation inconclusive state :([/url] Thanks for the report. Basically the inconclusive happens because the phone lost two stats units from the input (it should have crunched 9 stats units, but it only managed to return 7). As soon as something else completes the wu fully, this will get partial credit but obviously this is not optimal situation. I've seen this behavior on my own test Nexus and seems to especially get triggered when BOINC Client or the OS needs to suspend/kill/restart our app or BOINC Client during crunching. Everything seems to go more smoothly when I leave the phone to just crunch (and not do anything else with it) for the duration of the wu. That's not how people normally use their phones, though. :) There might be something wrong in our Android wrapper so that it doesn't cope with various oddities that running flakily (by design) on a phone brings. That's why I've left this as a beta app for now as our wrapper or Dnet Client settings might need more tweaks for Android. Also providing even smaller wu for Androids (with only one stat unit instead of the 9 our tiny wus have) might help as this ~5.5h crunch time is a bit long. -w
71) Message boards : Number crunching : ATI all work errors (Message 6211) Posted 29 Sep 2014 by Teemu Mannermaa Post: [Sep 21 20:51:08 UTC] Unable to initialize OpenCL. [Sep 21 20:51:08 UTC] Break Shutting down... ... 13:51:08 (5184): premature exit detected, app exit status: 0xfffffffd Thanks for the report! I've seen few hosts failing with this message. The error comes from the Distributed.net Client we call to do the actual crunching which for some unknown reason fails to initialize OpenCL. I'll contact the upstream devs to see if we can improve the error message or see if they can figure out what's going wrong on these hosts. As a first step, can you check if you can find any errors from the Windows event logs around the time BOINC Client tried to run our OpenCL app? Also this seems to be detected as a Win 8.1 core edition? I didn't know desktop side came with those stripped down core editions or is this in reality a Windows Server 2012 R2? Looks like you are using (old) development version of BOINC Client. Any reason for that? Can you try to use stable release (v7.2.42 at the moment) or at least update to the latest development version (v7.4.22 at the moment)? Is it possible for you to obtain a clinfo output from the system? AMD OpenCL drivers should install a clinfo binary in the system directory so it should be in PATH and you can just run it on any command line prompt. We might also be interested to see how Distributed.net Client detects your OS and cards. Can you change to the moowrap.net project directory (inside the BOINC Client data directory) and run command dnetc520-win32-x86-opencl.exe -cpuinfo -l \dnetc-opencl.log and then send the contents of \dnetc-opencl.log log file to us? It's okay if you don't feel confident working on the command line but otherwise feel free to ask more specific instructions about these two debug commands. Thank you! -w
72) Message boards : News : New app v1.4 with OpenCL support deployed (Message 6210) Posted 29 Sep 2014 by Teemu Mannermaa Post: I tested the new app on different platforms and everything went well. That's good to hear! - A way to select between opencl and ati/cal/stream (stream is much better for HD6xxxx cards) The way BOINC Server works is that it sends both variants and if one of them is constantly faster, it should start preferring that variant. It will still occasionally send the other one to see if it has gotten any better. I might add a project profile setting that allows one to disable OpenCL or native variant. - A way to choose the right core for opencl applications in the preferences pages. For instance the default (auto) selected core (core #1 CL 1-pipe) is slower than core #3 (according to benchmark results) on my 290-X. At the moment the AMD/ATI OpenCL app uses the ATI Stream setting (just as nVidia OpenCL app will use nVidia CUDA setting) to override the detected core for the app. Do you need a separate override for each type or is it enough for now? I might add two OpenCL specific core overrides for the next app version. -w
73) Message boards : News : New app v1.4 with OpenCL support deployed (Message 6175) Posted 9 Sep 2014 by Teemu Mannermaa Post: The Parameters in the line "LLVM ERROR: Cannot select:" are different in every Job (eg.: http://moowrap.net/result.php?resultid=29696345): Hmm, according to http://devgurus.amd.com/thread/160579 that seems to be a bug in the driver. According to your driver version 1084.4 you are using Catalyst 13.1 drivers, is that correct? They claim the bug is fixed in Catalyst 13.4 or newer. Can you try upgrading your driver and see what happens? -w
74) Message boards : News : New app v1.4 with OpenCL support deployed (Message 6172) Posted 9 Sep 2014 by Teemu Mannermaa Post: I would like to choose between ATI stream and OpenCL, but I can't see any way to do it. W7 64bit, ATI R9 280, driver: 1.4.1848 OpenCL: 1.2, CPU X6 1090T, Boinc core 7.2.42. OpenCL application makes 67 % GPU load only, rising a priority of the CPU process has no effect. Run times are 1,3x longer compared to 1.03 ATI stream application. If the run times are constantly slower, the BOINC Server should start favoring the other app for you after a while. At the moment there's no setting to force using only specific app but I might add it in the future. It's possible you need to play with core settings (they are in project preferences) to select more optimal one to tax the GPU more. You can run Dnet Client from the project dir with "dnetc520-win32-x86-opencl.exe -l \output.log -bench rc572" command and read through the log file to see what it recommends for you. I should somehow automate grabbing that benchmark or preferably the whole thing. App_info.xml doesn't work, disabled on the server side? The app_info.xml is fully on the client side, server doesn't have anything to say for it so it's not disabled. If you only have ATI Stream (grab the new versions, though) in it, you should only get jobs for that. -w
75) Message boards : News : New app v1.4 with OpenCL support deployed (Message 6171) Posted 9 Sep 2014 by Teemu Mannermaa Post: ...and it's the first GPU project that really seems to be taxing my GPU, as the fans kick in to overdrive as compared to Einstein, SETI, etc. That's a good sign for the GPU load. :) I used to use the Win 7 widget "GPU meter" from Addagadget.com on my old HD4670, though none of their versions seem to work with the new card so I don't know what the utilization is like. The old HD4670 would hover around 90% I use GPU-Z to see what my GPU utilization is and usually when GPUs are crunching away they hover around 97% for both of my cards. Normally I only watch that my temps are highish so that at least something is crunching. :) -w
76) Message boards : News : New app v1.4 with OpenCL support deployed (Message 6157) Posted 7 Sep 2014 by Teemu Mannermaa Post: Finally we have a newer application version that supports OpenCL! Updated versions of both AMD/ATI Stream and OpenCL apps for Windows have been offered by our server since yesterday to participants meeting requirements. These apps require BOINC Client v7.2.28 or newer and Windows XP or newer. The Stream app is not available to AMD Radeon HD 7870/7950/7970/R9 280X series (Tahiti) or newer cards due to compatibility issues. OpenCL app is preferred for those cards. The old app version is still available for systems running older BOINC Client. However, if your system has multiple GPUs, please consider updating it's BOINC Client to a support version. There are known issues running the old app on systems with multiple GPUs. New apps fix these by running one cruncher per device. For detailed app change history, please read application changelog. For latest requirements enforced by our server, please read latest requirements. Please report any problems you might encounter with new apps and our changes. Especially if you no longer get any work from our server where you previously did. You can report them by posting at our forum. Happy crunching!
77) Message boards : Number crunching : Dual+ GPUs (Problems and possible solutions) (Message 6154) Posted 5 Sep 2014 by Teemu Mannermaa Post: Just one thing: Do you have any benchmarks of the OpenCL application running on pre-Tahiti AMD/ATI GPUS (5xxx, 6xxx)? If there will be a significant drop in performance it may be worthy to also keep the ATI/CAL applications alive (the one based on dnetc518-win32-x86-stream, maybe adding the -devicenum to its runtime flags) I don't have anything conclusive about performance differences at the moment. OpenCL is supposed to be slightly slower but Dnet Client devs (and I'm sure card driver devs) have been improving OpenCL performance steadily. My own 5870 tests about the same on both OpenCL and Stream but it might not be able to use it's best core due to those getting disabled on my test system (because those cores would crash due to my 7970). For now I'm keeping Stream apps alive but we'll see what happens in the long run. Deploying with newer wrapper version (and Dnet Client one since build 519 is first one to support -devicenum option) is something I'd want to do for them to get new version benefits. -w
78) Message boards : Number crunching : Any chance to get support of R9 290X on this project ? (Message 6153) Posted 5 Sep 2014 by Teemu Mannermaa Post: Since the server move is over, can we get our OpenCL application for modern ATI GPUs ? Yes, we can and that's what I've been working for past couple of weeks. I just deployed beta version of OpenCL for ATI app. You need to opt in for beta apps in your preferences if you want to test it. -w
79) Message boards : Number crunching : Dual+ GPUs (Problems and possible solutions) (Message 6151) Posted 5 Sep 2014 by Teemu Mannermaa Post: this addition, it will ever use device 0. But if it is started with the -devicenum <n> command line argument (run on device <n> only), all the problem related to this issue will be solved. This should be just a minor change on the wrapper code. It does seem a minor change but it's complicated by the factor that BOINC Client gives apps 3 different ways the device number to use. And most of the time it gives us wrong GPU information (not for the one app should use, but for the other one). This is sad especially since the client knows the info, it just doesn't pass it via it's API. I've also seen the multi-GPU scheduling hang. I think that's a BOINC Client bug where they no longer schedule multiple GPUs correctly. :( However, your three problems are valid ones and summary the situation nicely. I'm happy to report that the upcoming OpenCL app will make use of the -devicenum to run one instance per device. It also fetches necessary GPU information directly from the BOINC Client by hooking into the process based detection feature in recent versions. That app version will hit beta any minute now. :) -w
80) Message boards : News : We have badges! (Message 6147) Posted 3 Sep 2014 by Teemu Mannermaa Post: However, don't you believe it would be nicer to have applications working properly on all GPU's (i.e. without app_info etc for 7950,7970, R9, ...)rather than badges ? That's indeed "nicer", that's why the badges are the stock ones as they were easy to just enable. :) Even if changing badges and implementing new/different ones is not top priority, it's good to have recommendations/suggestions bouncing around. At the moment I'm working on getting the wrapper refreshed so I can deploy the newer Dnet Clients that should fix various problems. Including getting an OpenCL app out that should work much better on newer cards. At least stability wise, if not necessarily performance. -w

Previous 20 · Next 20