Hi Bogdan, Fred, all, A few comments:
On Tue, 21 Jun 2011 01:29:57 +1000, Bogdan Opanchuk <manti...@gmail.com> wrote: > And now I see where my problem with understanding current architecture > is: I had somewhat different vision in mind. Compyte was supposed to > contain Python-wrapped Cuda/CL kernels (pyfft, randoms, scan, reduce > etc) and (ndarray-like) array class. That sounds about like what I think compyte might be. > PyCUDA and PyOpenCL are providing > low-level access to Cuda/CL functions (well, that's what they do now); > GPUArray is moved from PyCUDA to compyte The implementation does move, but the programming interfaces will stay where they are. > and generalized to work with PyOpenCL too. Right--in fact, PyOpenCL's current array is slightly cleaner (and a bit more mature) than PyCUDA's. > Compyte is a separate package and has its own setup > script; Again, I don't necessarily agree here. Andreas
pgpPkieqoM5CP.pgp
Description: PGP signature
_______________________________________________ PyCUDA mailing list PyCUDA@tiker.net http://lists.tiker.net/listinfo/pycuda