CUBLAS cublasCgemm use

Hi all,

is the following syntax correct?

cublasCgemm (‘N’,‘N’, N1_pad, M_pad, M_pad, alpha, Kernel_y_pad, N1_pad, DATA_pad, M_pad, beta, res_y, N1_pad);

where Kernel_y_pad is a N1_padXM_pad (rowsXcols) matrix and DATA_pad is a M_padxM_pad matrix.

It should work properly and provide a matrix res_y N1_pad X M_pad sized.

The results is very strange and not correct.

alpha and beta are cuFloatComplex equal to 1 and 0, respectively.

Thank you,

P