i have a quick question about the dimension of the result by mkl_dscsmm
i am doing, C = A'*B
and A is a m * n sparse matrix, and B is a m*k matrix, so i call the function like this,
alpha = 1;
beta = 0;
mkl_dcscmm("T", &m, &k, &n, &alpha, "GXXC", Ax, Ai, Ab, Ae, B, &k, &beta, C, &k);
is there any problem in my function call?
it looks that i can get the right result on when k = 1, for any other k != 1, i get trash. Do i do something totally wrong? any help is appreciated.