The Intel Math Kernel Library (MKL) contains many routines that can be linked to both Fortran and C/C++ programs. The Intel MKL enhances performance of the application programs that use it because the library has been optimized for the latest generations of Intel processors.

Among the available routines are BLAS, LAPACK, ScaLAPACK, Sparse Solvers, Vector Mathematical Functions, Vector Generators of Statistical Distributions, Discrete Fourier Transform Functions, and Fast Fourier Transforms. They are located in /opt/intel/mkl/9.0/lib.

First source the appropriate file:

• bash: source /opt/intel/mkl/9.0/tools/environment/mklvars32.sh
• tcsh: source  /opt/intel/mkl/9.0/tools/environment/mklvars32.csh

Typically you will need to link libmkl_ia32.a, libmkl_lapack.a, libguide.a and /usr/lib/libpthread.so. If you are using the sparse solver routines, you will also need to link libmkl_solver.a. See the Intel documentation /opt/intel/mkl/9.0/doc/Getting_Started.htm#Linking_Your_Application_with_Intel_MKL.

Online Documentation

• Getting Started: http://www.intel.com/software/products/mkl/docs/mklgs_lnx.htm
  CTC linux login machines: /opt/intel/mkl/9.0/doc/Getting_Started.htm
• Reference Manual: http://www.intel.com/software/products/mkl/docs/mklman.htm
  CTC linux login machines: /opt/intel/mkl/9.0/doc/mklman.pdf
• Quick Reference: http://www.intel.com/software/products/mkl/docs/mklqref/index.htm
  CTC linux login machines: /opt/intel/mkl/9.0/doc/mklqref/index.htm
• Technical User Notes: http://www.intel.com/software/products/mkl/docs/mklusew.htm
  CTC linux login machines: /opt/intel/mkl/9.0/doc/mkluse.htm