Hi All, As a few of you may remember, back in June we announced PyCULA<http://www.mail-archive.com/pycuda@tiker.net/msg01236.html>, a complete set of python bindings for CULAtools LAPACK that integrates with PyCUDA. A short time later in August our software was recognized by CULA<http://www.culatools.com/blog/2010/08/03/pycula-–-python-bindings-for-cula/>. Now I am pleased to announce our first official public release of PyCULA!
PyCULA is available from: http://pypi.python.org/pypi/PyCULA/0.9a4 Main Features Include: - ctypes/numpy bindings for LAPACK routines supported in CULA - PyCUDA/GPUArray interface for CULA device functions - *Mixing PyCUDA kernel code with LAPACK calls* - Allows for custom pitching/padding on device routines - Automatic memory management ala PyCUDA/Python - Develop, debug, and get help interactively in the interpreter Currently we support CULA2.1Premium in Single and Double precision on Linux x86_64 with CUDA3.1 and the most recent PyCUDA. Other versions of CULA including CULA basic should work with respect to their limited features. In our next release, we hope to add Mac64 support, enhanced context management, CUDA3.2 support, as well as optimize some of our code. I hope that our package will be useful to some of you, especially the mixing of kernel and runtime code, which is what inspired PyCULA. Garrett @PyCULA garrett.wri...@temple.edu
_______________________________________________ PyCUDA mailing list PyCUDA@tiker.net http://lists.tiker.net/listinfo/pycuda
