Full matrix storage in MKL LAPACK

[topol555]

Please, help me. I'm using dgels function (declared in mkl_lapack.h) from MKL LAPACK package in my program on C++. I can't understand how should I store my full matrix in one-dimensional array - in column-major or row-major format.

For example, if matrix A=
(1.0 2.0)
(3.0 4.0)
(5.0 6.0)
then the array should be
1.1) [1.0, 2.0, 3.0, 4.0, 5.0, 6.0] (row-major format) or
1.2) [1.0, 3.0, 5.0, 2.0, 4.0, 6.0] (column-major format).

And if Ax=b, where b is a vector (7.0, 8.0, 9.0), then b will be overwritten by the solution vector like
2.1) [x1, x2, 9.0] or
2.2) [x1, 8.0, x2].

[TimP]

In case you didn't find answers in MKL docs,
http://www.netlib.org/clapack/readme
describes the storage from the C point of view (column-major)
I think you would have to specify more about your usage to understand the second part of your question. However, in the cases where the result doesn't use all the entries of b, it is the leading entries which are over-written.

[Gennady_F_Intel]

Quoting - topol555

Please, help me. I'm using dgels function (declared in mkl_lapack.h) from MKL LAPACK package in my program on C++. I can't understand how should I store my full matrix in one-dimensional array - in column-major or row-major format.

For example, if matrix A=
(1.0 2.0)
(3.0 4.0)
(5.0 6.0)
then the array should be
1.1) [1.0, 2.0, 3.0, 4.0, 5.0, 6.0] (row-major format) or
1.2) [1.0, 3.0, 5.0, 2.0, 4.0, 6.0] (column-major format).

And if Ax=b, where b is a vector (7.0, 8.0, 9.0), then b will be overwritten by the solution vector like
2.1) [x1, x2, 9.0] or
2.2) [x1, 8.0, x2].

Please look at the following link:
ttp://software.intel.com/sites/products/documentation/hpc/mkl/lapack/mkl_lapack_examples/index.htm

you can find there the C example of DGELS routine.

/* Solve the equations A*X = B */
dgels( "No transpose", &m, &n, &nrhs, a, &lda, b, &ldb, work, &lwork, &info );
--Gennady

[topol555]

Thank you!