I've been trying to get PyOpenCL and PyCUDA running on a Linux Mint machine. I have things installed but the demo scripts fail with the error:
pyopencl.cffi_cl.LogicError: clgetplatformids failed: PLATFORM_NOT_FOUND_KHR
Configuration
$ uname -a && cat /etc/lsb-release && lspci | grep NV
Linux 3.13.0-24-generic #47-Ubuntu SMP Fri May 2 23:30:00 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux
DISTRIB_DESCRIPTION="Linux Mint 17.3 Rosa"
01:00.0 VGA compatible controller: NVIDIA Corporation GK208 [GeForce GT 730] (rev a1)
Relevant installed packages:
libcuda1-352-updates
libcudart5.5:amd64
nvidia-352-updates
nvidia-352-updates-dev
nvidia-cuda-dev
nvidia-cuda-toolkit
nvidia-opencl-icd-352-updates
nvidia-profiler
nvidia-settings
ocl-icd-libopencl1:amd64
ocl-icd-opencl-dev:amd64
opencl-headers
python-pycuda
python-pyopencl
python3-pycuda
python3-pyopencl
Research
This post describes a scenario in which the package-manager installed opencl/cuda implementation don't set up some simlinks correctly. That issue doesn't seem to be present on my system.
There was a version number mismatch between the graphics drivers (were nvidia-340) and the nvidia-opencl package (352). I update the graphics drivers to nvidia-352-updates-dev but the issue remains.
There is a bug in Arch linux that seems to revolve around the necessary device files not being created. However, I've verified that the /dev/nvidia0 and /dev/nvidiactl exist and have permissions 666, so they should be accessible.
Another Stackoverflow post suggests running the demos as root. I have tried this and the behavior does not change.
Older installation instructions for cuda/opencl say to download drivers directly from the NVidia website. I'm not sure this still applies, so I'm holding off on that for now since there seem to be plenty of relevant packages in the respositories.
The same error, but for an ATI card on a different linux system, was resolved by putting proper files in /usr/lib/OpenCL/vendors. That path isn't used on my system, However, I do have /etc/OpenCL/vendors/nvidia.icd which contains the line libnvidia-opencl.so.1, suggesting my issue is dissimilar.
This error has been observed on OSX, but for unrelated reasons. Similar error messages for PyCUDA also appear to be unrelated.
This error can occur under remote access since the device files are not initialized if X is not loaded. However, I'm testing this in a desktop environment. Furthermore, I've run the manual commands suggested in that thread just to be sure, and they are redundant since the relevant /dev entries already exist.
There is a note here about simply running a command a few times to get around some sort temporary glitch. That doesn't seem to help.
This post describes how the similar cuInit failed: no device CUDA error was caused by not having the user in the video group. To check, I ran usermod -a -G video $USER, but it did not resolve my issue.
In the past, routine updates have broken CUDA support. I have not taken the time to explore every permutation of package version numbers, and it's possible that downgrading some packages may change the situation. However, without further intuition about the source of the issue, I'm not going to invest time in doing that since I don't know whether it will work.
The most common google search result for this error, appearing four times on the first pages, is a short and unresolved email thread on the PyOpenCL list. Checking the permissions bits for /dev/nvidia0 and /dev/nvidiactl is suggested. On my machine user/group/other all have read and write access to these devices, so I don't think that's the source of the trouble.
I've also tried building and installing PyOpenCL form the latest source, rather than using the version in the repositories. This is failing at an earlier phase which suggests to me it is not building correctly.
Summary
The issue would appear to be that PyCUDA/PyOpenCL cannot locate the graphics card. There are several known issues that can cause this, but none of them seem to apply here. I'm missing something, and I'm not sure what else to do.
I'm trying to build a VRPN server with Python3 flag using Python 3.4 64-bit on Windows 7 64-bit but there seems to be a problem. I need this for BlenderVR software.
This is my procedure:
1) I use CMake to create makefiles (I'm using 3.4.0 version but I've also tried different ones). I do it with this command (those flags should be there but the result seems to be the same without them anyway):
cmake -G"MinGW Makefiles" -HD:\My\BlenderVR\plugins\vrpn
-BD:\My\BlenderVR\plugins\cmake -DVRPN_BUILD_PYTHON=OFF -DVRPN_BUILD_PYTHON_HANDCODED_2X=OFF -DVRPN_BUILD_PYTHON_HANDCODED_3X=ON
I used to add those flags as well but it seems that it can find Python without them
-DPYTHON_INCLUDE_DIR=D:\My\BlenderVR\Required\Python3\include
-DPYTHON_LIBRARY=D:\My\BlenderVR\Required\Python3\libs\python34.lib
Python is correctly found and this operation doesn't throw any error.
2) Then I use mingw32-make.exe to build it and I get this error:
[ 90%] Linking CXX shared module vrpn.pyd D:/My/BlenderVR/Required/Python3/libs/python34.lib: error adding
symbols: File f ormat not recognized collect2.exe: error: ld
returned 1 exit status
python\CMakeFiles\vrpn-python.dir\build.make:505: recipe for
target 'python/vrpn .pyd' failed mingw32-make[2]: * * *
[python/vrpn.pyd] Error 1 CMakeFiles\Makefile2:3247: recipe for
target 'python/CMakeFiles/vrpn-python.dir/ all' failed
mingw32-make[1]: * * * [python/CMakeFiles/vrpn-python.dir/all]
Error 2 Makefile:159: recipe for target 'all' failed
mingw32-make: [all] Error 2
vprn.pyd is the crucial thing for my future work.
I figured out that it needs libpython34.a file (probably). When I created it and copied to Python3/libs folder it worked and finished without errors but the crated vprn.pyd didn't worked as it should.
What I need is to get import vrpn to work with this simple test in python (appending path where vrpn.pyd was build):
import sys
sys.path.append('D:/My/BlenderVR/plugins/cmake/python')
import vrpn
It lags my whole computer for a while and then pops out that Python has stop working.
I suspect that problem is in the libpython34.a file that I created doing this:
gendef python34.dll (in Windows/System32)
dlltool -D python34.dll -d python34.def -l libpython34.a
I don't how else should I get the libpython file. I've tried various versions of CMake and MinGW (like MinGWPy, TDM, w64) with many CMake flags. I was able to make it work using 32-bit Python but I need 64-bit version otherwise it is not working with BlenderVR enviroment.
I know this is very specific problem and probably kind of confusing at first but I didn't know how else to put it. I'll be glad for anything that could help. Thank you.
mingwpy should be installed with pip (until it is officially released at PYPI):
pip install -i https://pypi.anaconda.org/carlkl/simple mingwpy
all necessary import files are atomatically copied into the python\libs folder.
If python\Scripts is in the PATH it should work out of the box.
You have to make sure, that Blender Python is equiped with two import files
D:\My\BlenderVR\Required\Python3\libs\libpython\libpython34.dll.a
D:\My\BlenderVR\Required\Python3\libs\libpython\libmsvcr100.a
I'm exploring the Qt framework and I'm learning about debugging a Qt application.
I've set a breakpoint on a line in the source code and click the "Start Debugging" icon/button. The console prints "Debugging starts"...and then a Dialog Box appears stating:
"Python quit unexpectedly while using _lldb.so plug-in".
Then the whole program just runs without stopping at the breakpoint I've set.
How do I fix this problem with Qt?
Thank you...
I'm using a MacBookPro running on OS X Yosemite.
I'm using:
Qt Creator 3.3.0 (opensource)
Based on Qt 5.4.0 (Clang 6.0 (Apple), 64 bit)
Built on Dec 8 2014 at 15:34:58
From revision d36c4d87db
Copyright 2008-2014 Digia Plc. All rights reserved.
How do I fix this?
Thank you
The shared library (.so) provided with the QT download unfortunately incompatible between Mavericks and Yosemite.
To fix this you need to compile Qt by yourself. Just follow these steps (you might have to adapt a little bit depending on your original Qt install):
git clone https://gitorious.org/qt-creator/qt-creator.git
cd qt-creator
export PATH=${PATH}:/Users/${user}/Qt/5.4/clang_64/bin/
qmake
make
cd bin/Qt\ Creator.app/Contents/MacOS
./Qt\ Creator
Hope it works for you.
This is probably https://bugreports.qt.io/browse/QTCREATORBUG-13803 which is fixed in the meantime.
Compiling Xcode Project fails with following errors:
'missing required architecture arm64 in file /Users/*/Git/ocr/opencv2.framework/opencv2'
It works well, if i change Architectures(under Build Settings) to (armv7, armv7s) instead of (armv7, armv7s).
How to change the opencv python build script, to add arm64 support to opencv2.framework?
The latest OpenCV iOS framework supports 64 bit by default
It can be downloaded at: OpenCV download page
I modified the following to make it build, though I haven't got an arm64 iOS device to test at the moment.
Edit: I also had to follow https://stackoverflow.com/a/17025423/1094400
Assuming "opencv" is the folder containing the opencv source from Github:
in each of gzlib.c, gzread.c, gzwrite.c located in opencv/3rdparty/zlib/ add:
#include <unistd.h>
at the top after the existing include.
In addition open opencv/platforms/ios/cmake/Modules/Platform/iOS.cmake and change line 88 from:
set (CMAKE_OSX_ARCHITECTURES "$(ARCHS_STANDARD_32_BIT)" CACHE string "Build architecture for iOS")
to:
set (CMAKE_OSX_ARCHITECTURES "$(ARCHS_STANDARD_INCLUDING_64_BIT)" CACHE string "Build architecture for iOS")
Furthermore change the buildscript at opencv/platforms/ios/build_framework.py in lines 99 and 100 from:
targets = ["iPhoneOS", "iPhoneOS", "iPhoneSimulator"]
archs = ["armv7", "armv7s", "i386"]
to:
targets = ["iPhoneOS", "iPhoneOS", "iPhoneOS", "iPhoneSimulator", "iPhoneSimulator"]
archs = ["armv7", "armv7s", "arm64", "i386", "x86_64"]
The resulting library will include the following:
$ xcrun -sdk iphoneos lipo -info opencv2
Architectures in the fat file: opencv2 are: armv7 armv7s i386 x86_64 arm64
Although I have a remaining concern regarding opencv/platforms/ios/cmake/Toolchain-iPhoneOS_Xcode.cmake which defines the size of a data pointer to be 4 in lines 14 and 17.
It should be 8 for 64bit I guess, so as I haven't tested if the compiled library is working for arm64 I would suggest further investigations at this point if it does not run properly.
micahp's answer was almost perfect, but missed the simulator version. So modify platforms/ios/build_framework.py to:
targets = ["iPhoneOS", "iPhoneOS", "iPhoneOS", "iPhoneSimulator", "iPhoneSimulator"]
archs = ["armv7", "armv7s", "arm64", "i386", "x86_64"]
You'll need to download the command line tools for Xcode 5.0.1 and then run
python opencv/platforms/ios/build_framework.py ios
Try to wait a next month. Will release a new XCode with more powerful supporting of 32/64 bit.
https://developer.apple.com/news/index.php?id=9162013a
Modify "build_frameworks.py" to:
def build_framework(srcroot, dstroot):
"main function to do all the work"
targets = ["iPhoneOS", "iPhoneOS", "iPhoneOS", "iPhoneSimulator"]
archs = ["armv7", "armv7s", "arm64", "i386"]
for i in range(len(targets)):
build_opencv(srcroot, os.path.join(dstroot, "build"), targets[i], archs[i])
put_framework_together(srcroot, dstroot)
#Jan, I followed your instructions, but OpenCV still doesn't run on arm64. You made such a detailed and wonderful answer - why not check it out on a simulator and see if you can make it run? :-)
FWIW, I think it might be harder than it seems. On the openCV stackoverflow clone, there's an indication that this problem might be non-trivial.
Instead of using terminal commands given in the opencv installation guide in official website, use the following commands. Worked for me.
cd OpenCV-2.3.1
mkdir build
cd build
cmake -G "Unix Makefiles" ..
make
sudo make install
I was having a similar error, but the issue wasn't related with the arm64 coompilation.fixed adding the framework libc++.dylib
Hi I'm having trouble installing and using the Mercurial ChartExtension
When I installed it as per instructions
First you need to install the extension; type this in your shell:
python ./setup.py install
Blockquote
I then modified my mercurial.ini file as follows
[extensions]
chart=/path/to/chart.py
Blockquote
Then tried running Hg Chart command and got the following error
Can anyone help me get this extension working. I know there's the Hg ActivityExtension as well, but i have not had much luck with that either Problem installing Mercurial Activity extension
A layman's guide to what steps i need to follow would be of immense help
Based on the changelog, the Chart extension has not been updated since late 2008. It would probably work if you tried using it with a version of Mercurial released around that same time (Mercurial v1.1.2 was released on Dec. 30, 2008).
The latest version of Mercurial is now v1.9.2. The API has changed (probably quite a bit) since v1.1.2. If the extension has not been modified to keep up with the changes to the API, then it will fail in ways similar to the error you found. In this case, the number of arguments for the walkchangerevs method has changed.
Updating the extension could be a large task...there is no way to know without inspecting the code. You could try to contact the author (#Ry4an) and ask for help. You could also try to modify the extension yourself.