http://permalink.gmane.org/gmane.comp.python.cuda hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://permalink.gmane.org/gmane.comp.python.cuda/2910 <pre> Done. Thanks! Andreas </pre> Andreas Kloeckner 2012-05-20T19:37:44 http://permalink.gmane.org/gmane.comp.python.cuda/2909 <pre>Dnia 2012-04-22, nie o godzinie 16:51 -0400, Andreas Kloeckner pisze: I have renamed _pycuda_struct.cpp to _pvt_struct_v2.cpp adding ifdef, added _pvt_struct_v3.cpp, and copied snippets from PyOpenCL setup.py. It builds on Debian now. Lintian (Debian tool checking package validity) complains a bit, but at least I have both Python2 and Python3 binaries. I also had to change call to hack_distutils because calling it with debug=True caused build to fail. See attached patch. I had to fix tests to run on Python3 (see patch). Almost all fail now because of Unicode hashing, but at least they run, so we can start fixing them. Andreas, can you apply those two patches? Best regards. </pre> Tomasz Rybak 2012-05-20T19:05:25 http://permalink.gmane.org/gmane.comp.python.cuda/2908 <pre> Try building with CC=gcc python setup.py build ? (shot in the dark) </pre> Andreas Kloeckner 2012-05-18T01:12:54 http://permalink.gmane.org/gmane.comp.python.cuda/2907 <pre>hi all, we have been trying, unsuccessfully to get pycuda installed on a new mac pro with a quadro fx 4800. we have nvida toolkit and drivers installed correctly and can build the test binaries just fine. our error/log messages are below - if anyone has suggestions we would be most appreciative. thanks in advance. </pre> Serge Rey 2012-05-17T23:28:33 http://permalink.gmane.org/gmane.comp.python.cuda/2906 <pre>LinkedIn ------------ Saito Norio requested to add you as a connection on LinkedIn: ------------------------------------------ Michael, I'd like to add you to my professional network on LinkedIn. - Saito Accept invitation from Saito Norio http://www.linkedin.com/e/-s1x6c6-h2adwi65-t/vBkXylyN9hJL27A5s4uIp6ISq9x/blk/I101001497_45/1BpC5vrmRLoRZcjkkZt5YCpnlOt3RApnhMpmdzgmhxrSNBszYRd5YTejgNc30Nc359bShDs6tPpP91bPcSdPwQczANcjsLrCBxbOYWrSlI/EML_comm_afe/?hs=false&amp;tok=1-evilL1t-3Rg1 View invitation from Saito Norio http://www.linkedin.com/e/-s1x6c6-h2adwi65-t/vBkXylyN9hJL27A5s4uIp6ISq9x/blk/I101001497_45/3kQnPsVd34Mc34MckALqnpPbOYWrSlI/svi/?hs=false&amp;tok=1yVbKbycR-3Rg1 ------------------------------------------ Why might connecting with Saito Norio be a good idea? Saito Norio's connections could be useful to you: After accepting Saito Norio's invitation, check Saito Norio's connections to see who else you may know and who you might want an introduction to. Building these connections can create opp</pre> Saito Norio via LinkedIn 2012-05-16T12:46:07 http://permalink.gmane.org/gmane.comp.python.cuda/2905 <pre>So, i need to use shared memory and synchronization in order to avoid it? Or it doesn't matter? Thank you! -- View this message in context: http://pycuda.2962900.n2.nabble.com/different-results-nan-values-tp7556797p7559569.html Sent from the PyCuda mailing list archive at Nabble.com. </pre> ggeo 2012-05-15T09:42:15 http://permalink.gmane.org/gmane.comp.python.cuda/2904 <pre> Easy to test: does it run fine single-threaded? :) Also, if you're using the prefork mechanism, that's definitely not thread-safe, because all threads share a pipe to communicate with the pre-forked nvcc runner. Andreas </pre> Andreas Kloeckner 2012-05-14T23:38:53 http://permalink.gmane.org/gmane.comp.python.cuda/2903 <pre>I've seen this error a few times, but it's not reproducible. Can anyone give any insight into what might be going wrong? Traceback (most recent call last): File "/home/elis/edit/work/dev/mms/common/util/threads.py", line 219, in run mod = cudahelper.compileSourceModule(kernel.code_str, kernel.buildOptions_list) File "/home/elis/edit/work/dev/mms/common/util/cudahelper.py", line 551, in compileSourceModule return SourceModule(const_src + textwrap.dedent(source), options=list(default_build_options | set(build_options)), nvcc=os.path.join(os.getenv('CUDA_HOME', '/usr/local/cuda'), 'bin', 'nvcc'), **kwargs) File "/home/elis/venv/dev/local/lib/python2.7/site-packages/pycuda-2011.2.2-py2.7-linux-x86_64.egg/pycuda/compiler.py", line 286, in __init__ self.module = module_from_buffer(cubin) LaunchError: cuModuleLoadDataEx failed: launch failed - This happened in some (but not all, I don't think) of the threads that got launched at roughly the same time. A quick google search didn't turn up anyth</pre> Eli Stevens (Gmail 2012-05-14T22:20:41 http://permalink.gmane.org/gmane.comp.python.cuda/2902 <pre> Welcome to parallel programming. https://en.wikipedia.org/wiki/Race_condition :) Andreas </pre> Andreas Kloeckner 2012-05-14T16:55:45 http://permalink.gmane.org/gmane.comp.python.cuda/2901 <pre>Hello, I am running a program which ,generally,runs fine but, when i am taking the result ( a matrix) sometimes it's ok and sometimes it has nan values inside. I can't understand that behaviour. (i am using linux 64bit) -- View this message in context: http://pycuda.2962900.n2.nabble.com/different-results-nan-values-tp7556797.html Sent from the PyCuda mailing list archive at Nabble.com. </pre> ggeo 2012-05-14T16:14:59 http://permalink.gmane.org/gmane.comp.python.cuda/2900 <pre>I have it but at different location: ......&lt; at &gt;ubuntu:/usr$ locate libcuda.so /usr/lib/nvidia-current/libcuda.so /usr/lib/nvidia-current/libcuda.so.1 /usr/lib/nvidia-current/libcuda.so.295.40 /usr/lib/nvidia-current-updates/libcuda.so /usr/lib/nvidia-current-updates/libcuda.so.1 /usr/lib/nvidia-current-updates/libcuda.so.295.40 /usr/lib32/nvidia-current/libcuda.so /usr/lib32/nvidia-current/libcuda.so.1 /usr/lib32/nvidia-current/libcuda.so.295.40 /usr/lib32/nvidia-current-updates/libcuda.so /usr/lib32/nvidia-current-updates/libcuda.so.1 /usr/lib32/nvidia-current-updates/libcuda.so.295.40 I tried this, maybe you can figure something from it: ......&lt; at &gt;ubuntu:/usr/lib/nvidia-current$ file libcuda.so libcuda.so: symbolic link to `libcuda.so.295.40' ......&lt; at &gt;ubuntu:/usr/lib/nvidia-current$ file libcuda.so.295.40 libcuda.so.295.40: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV), dynamically linked, stripped ......&lt; at &gt;ubuntu:/usr/lib/nvidia-current$ and this: ......&lt; at &gt;ubuntu:/usr/lib32/nvidia-current$ file libcuda.s</pre> maxrider11 2012-05-11T04:18:12 http://permalink.gmane.org/gmane.comp.python.cuda/2899 <pre> Well, is there any libcuda.so* on your computer? Andreas </pre> Andreas Kloeckner 2012-05-11T03:02:55 http://permalink.gmane.org/gmane.comp.python.cuda/2898 <pre>ohh. ........&lt; at &gt;ubuntu:/usr/local/lib$ file /usr/local/lib/libcuda.so /usr/local/lib/libcuda.so: ERROR: cannot open `/usr/local/lib/libcuda.so' (No such file or directory) Then I tried this: ......&lt; at &gt;ubuntu:/usr/local/cuda/lib$ file /usr/local/cuda/lib/libcurand.so /usr/local/cuda/lib/libcurand.so: symbolic link to `libcurand.so.4' ......&lt; at &gt;ubuntu:/usr/local/cuda/lib$ file libcurand.so.4 libcurand.so.4: symbolic link to `libcurand.so.4.1.28' ......&lt; at &gt;ubuntu:/usr/local/cuda/lib$ file libcurand.so.4.1.28 libcurand.so.4.1.28: ELF *32-bit *LSB shared object, Intel 80386, version 1 (SYSV), dynamically linked, BuildID[sha1]=0xca8aacf0fff3da60e113fb2ede0e05a8ca3bd576, stripped does it help ? _______________________________________________ PyCUDA mailing list PyCUDA&lt; at &gt; http://lists.tiker.net/listinfo/pycuda -- View this message in context: http://pycuda.2962900.n2.nabble.com/installing-pycuda-2011-2-2-error-cannot-find-lboost-python-mt-tp7481143p7548995.html Sent from the PyCuda mailing list archive at Nabble.com. </pre> maxrider11 2012-05-11T02:59:19 http://permalink.gmane.org/gmane.comp.python.cuda/2897 <pre> Well, I meant: run 'file' on your libcuda.so. :) Andreas </pre> Andreas Kloeckner 2012-05-11T01:52:45 http://permalink.gmane.org/gmane.comp.python.cuda/2896 <pre>.....&lt; at &gt;ubuntu:~$ file /usr/lib/x86_64-linux-gnu/libOpenCL.so.1 /usr/lib/x86_64-linux-gnu/libOpenCL.so.1: ERROR: cannot open `/usr/lib/x86_64-linux-gnu/libOpenCL.so.1' (No such file or directory) -- View this message in context: http://pycuda.2962900.n2.nabble.com/installing-pycuda-2011-2-2-error-cannot-find-lboost-python-mt-tp7481143p7548821.html Sent from the PyCuda mailing list archive at Nabble.com. </pre> maxrider11 2012-05-11T00:44:32 http://permalink.gmane.org/gmane.comp.python.cuda/2895 <pre> Well, the linker said /usr/local/lib/libcuda.so is incompatible. Try this: $ file /usr/lib/x86_64-linux-gnu/libOpenCL.so.1 /usr/lib/x86_64-linux-gnu/libOpenCL.so.1: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV), dynamically linked, stripped This allows you to tell if it's a 32-bit library. Andreas </pre> Andreas Kloeckner 2012-05-11T00:10:17 http://permalink.gmane.org/gmane.comp.python.cuda/2894 <pre>Thank you for quick reply. someone else installed cuda. How to see that installed cuda is 32 bit ? If you are 100% sure about 32 bit, then I will install 64 bit cuda. -- View this message in context: http://pycuda.2962900.n2.nabble.com/installing-pycuda-2011-2-2-error-cannot-find-lboost-python-mt-tp7481143p7548693.html Sent from the PyCuda mailing list archive at Nabble.com. </pre> maxrider11 2012-05-10T23:27:33 http://permalink.gmane.org/gmane.comp.python.cuda/2893 <pre>Looks like you've installed a 32-bit CUDA on a 64-bit machine. Andreas On Thu, 10 May 2012 15:32:54 -0700 (PDT), maxrider11 &lt;ssethia86-Re5JQEeQqe8AvxtiuMwx3w&lt; at &gt;public.gmane.org&gt; wrote: </pre> Andreas Kloeckner 2012-05-10T22:53:19 http://permalink.gmane.org/gmane.comp.python.cuda/2892 <pre>py.so was py27.so which i edited later. output of failing command: *.......&lt; at &gt;ubuntu:~/pycuda-2011.2.2$ make -j 4* ctags -R src || true /usr/bin/python setup.py build running build running build_py running build_ext building '_driver' extension gcc -pthread -fno-strict-aliasing -fwrapv -Wall -g -fPIC -DPYGPU_PYCUDA=1 -DPYGPU_PACKAGE=pycuda -DHAVE_CURAND=1 -Isrc/cpp -I/usr/include -I/usr/local/cuda/include -I/usr/lib/python2.7/dist-packages/numpy/core/include -I/usr/include/python2.7 -c src/cpp/cuda.cpp -o build/temp.linux-x86_64-2.7/src/cpp/cuda.o gcc -pthread -fno-strict-aliasing -fwrapv -Wall -g -fPIC -DPYGPU_PYCUDA=1 -DPYGPU_PACKAGE=pycuda -DHAVE_CURAND=1 -Isrc/cpp -I/usr/include -I/usr/local/cuda/include -I/usr/lib/python2.7/dist-packages/numpy/core/include -I/usr/include/python2.7 -c src/cpp/bitlog.cpp -o build/temp.linux-x86_64-2.7/src/cpp/bitlog.o gcc -pthread -fno-strict-aliasing -fwrapv -Wall -g -fPIC -DPYGPU_PYCUDA=1 -DPYGPU_PACKAGE=pycuda -DHAVE_CURAND=1 -Isrc/cpp -I/usr/include -I/usr/local/cu</pre> maxrider11 2012-05-10T22:32:54 http://permalink.gmane.org/gmane.comp.python.cuda/2891 <pre> Not sure what -py.so is. Use -py27. No. 'ld' is the linker. Can you please post the output of the failing command? Andreas </pre> Andreas Kloeckner 2012-05-10T04:29:59 http://permalink.gmane.org/gmane.comp.python.cuda/2890 <pre>Hi, Sorry, I am late on followup. I am beginner and wandered too much, and fiddled a lot. I tried your advice, still same problem. This is what i did: *cuda installed *in /usr/local/cuda, it's version is 4.0. *calling configure.py *like this: $ ./configure.py --cuda-root=/usr/local/cuda --cudadrv-lib-dir=/usr/lib --boost-inc-dir=/usr/include --boost-lib-dir=/usr/lib --boost-python-libname=boost_python-mt-py27 --boost-thread-libname=boost_thread-mt I see these files in /usr/lib: libboost_python-py27.so and libboost_python-mt-py.so. so which to use ? and shall i add ".so" extension with their name when i call ./configure.py ? $ echo CUDA_ROOT prints usr/local/cuda/, then also it says "/usr/bin/ld: -lcuda not found". I don't see 'ld' directory in /usr/bin. I appreciate your help ! -- View this message in context: http://pycuda.2962900.n2.nabble.com/installing-pycuda-2011-2-2-error-cannot-find-lboost-python-mt-tp7481143p7545342.html Sent from the PyCuda mailing list archive at Nabble.com. </pre> maxrider11 2012-05-10T03:39:39 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.comp.python.cuda