Join GitHub today
cublasCgemmBatched, cublasZgemmBatched wrappers #52
I noticed that the wrappers for cublasCgemmBatched and cublasZgemmBatched were missing (the float and double versions are wrapped already). I need cublasCgemmBatched, so I added them.
I didn't write any tests for this since the modifications are pretty trivial (I just copied from cublasSgemmBatched and cublasCgemm), but I can have a go at writing one if it is necessary.
I am currently using scikits.cuda to build an FFT-based convolution operator for Theano. I need to do a batch of dot products in the Fourier domain to implement an input domain convolution, that's why I need cublasCgemmBatched. It's here, in case anyone is interested: https://github.com/benanne/theano_fftconv