intel math kernel library

[bekar__bahadır1]

I want to calculate the eigenvalue eigenvector fortranda.
I have a few questions here.
1- Select dynamic or static linking: what is dynamic static, single dynamic library
2- Select interface layer: 32 or 64
3- Select threading layer: open mpı squental ...?
4- Select OpenMP library: What is intel libiomp5md.lib
5- Select MPI library: What if I choose paralleling even if my program is not parallel?
6- Select the Fortran 95 interfaces: when should I check this
7- Use this link line: where do we use the links here?
8- Compiler options: where do we write this part?

[Steve_Lionel]

I suggest that you ask these questions in https://software.intel.com/en-us/forums/intel-math-kernel-library

[Gennady_F_Intel]

The answers on all of these questions you may find out into MKL Developer Guide :  https://software.intel.com/en-us/mkl-linux-developer-guide

[bekar__bahadır1]

i couldn't run the following sample code somehow

     .. Parameters ..
      INTEGER          N, NSELECT
      PARAMETER        ( N = 5, NSELECT = 3 )
      INTEGER          LDA, LDZ
      PARAMETER        ( LDA = N, LDZ = N )
      INTEGER          LWMAX
      PARAMETER        ( LWMAX = 1000 )
*
*     .. Local Scalars ..
      INTEGER          INFO, LWORK, IL, IU, M
      DOUBLE PRECISION ABSTOL, VL, VU
*
*     .. Local Arrays ..
*     IWORK dimension should be at least 5*N
      INTEGER          IFAIL( N ), IWORK( 5*N )
      DOUBLE PRECISION A( LDA, N ), W( N ), Z( LDZ, NSELECT ),
     $                 WORK( LWMAX )
      DATA             A/
     $  6.29, 0.00, 0.00, 0.00, 0.00,
     $ -0.39, 7.19, 0.00, 0.00, 0.00,
     $  0.61, 0.81, 5.48, 0.00, 0.00,
     $  1.18, 1.19,-3.13, 3.79, 0.00,
     $ -0.08,-0.08, 0.22,-0.26, 0.83
     $                  /
*
*     .. External Subroutines ..
      EXTERNAL         DSYEVX
      EXTERNAL         PRINT_MATRIX
*
*     .. Intrinsic Functions ..
      INTRINSIC        INT, MIN
*
*     .. Executable Statements ..
      WRITE(*,*)'DSYEVX Example Program Results'
*     Negative ABSTOL means using the default value
      ABSTOL = -1.0
*     Set IL, IU to compute NSELECT smallest eigenvalues
      IL = 1
      IU = NSELECT
*
*     Query the optimal workspace.
*
      LWORK = -1
      CALL DSYEVX( 'Vectors', 'Indices', 'Upper', N, A, LDA, VL, VU, IL,
     $             IU, ABSTOL, M, W, Z, LDZ, WORK, LWORK, IWORK, IFAIL,
     $             INFO )
      LWORK = MIN( LWMAX, INT( WORK( 1 ) ) )
*
*     Solve eigenproblem.
*
      CALL DSYEVX( 'Vectors', 'Indices', 'Upper', N, A, LDA, VL, VU, IL,
     $             IU, ABSTOL, M, W, Z, LDZ, WORK, LWORK, IWORK, IFAIL,
     $             INFO )
*
*     Check for convergence.
*
      IF( INFO.GT.0 ) THEN
         WRITE(*,*)'The algorithm failed to compute eigenvalues.'
         STOP
      END IF
*
*     Print the number of eigenvalues found.
*
      WRITE(*,'(/A,I2)')' The total number of eigenvalues found:', M
*
*     Print eigenvalues.
*
      CALL PRINT_MATRIX( 'Selected eigenvalues', 1, M, W, 1 )
*
*     Print eigenvectors.
*
      CALL PRINT_MATRIX( 'Selected eigenvectors (stored columnwise)',
     $                   N, M, Z, LDZ )
      STOP
      END
*
*     End of DSYEVX Example.
*
*  =============================================================================
*
*     Auxiliary routine: printing a matrix.
*
      SUBROUTINE PRINT_MATRIX( DESC, M, N, A, LDA )
      CHARACTER*(*)    DESC
      INTEGER          M, N, LDA
      DOUBLE PRECISION A( LDA, * )
*
      INTEGER          I, J
*
      WRITE(*,*)
      WRITE(*,*) DESC
      DO I = 1, M
         WRITE(*,9998) ( A( I, J ), J = 1, N )
      END DO
*
 9998 FORMAT( 11(:,1X,F6.2) )
      RETURN
      END

[Gennady_F_Intel]

This is exaclty the same example MKL provides in the form of examples. You need to check mklroot\examples\lapackf directory.

Check the makefile and learn how to build this example.