Yang,
thank you very much for your question.
MKL 7.0 does not have a routine for the covariance matrix calculation.
We are considering opportunities of the library functionality extensions.
Further development of MKL could take into account the covariance matrices.
We kindly appreciate to let us know about your expectations from this function.
What would you like to have as inputs of the routine (matrix of observations, dimensions of the matrix, etc)?
Would you like to have chance to determine frequencies and weights of observations, other observation control parameters?
What would be output (covariance matrix, variances, or partial covariances from the covariance matrix, etc)?
Everybody welcome to participate in discussions.
Thank you for your time,
Best regards,
Andrey
