LAPACK/BLAS

Basic Linear Algebra Subprograms (BLAS) library is an archive of common combination of Vector, and Matrix operations. Switch -llapack, and -lblas are usually used in the code compilation.

Library/Compiler Version

To check the version of Intel Compilers, type:

ifort --version

output:

ifort (IFORT) 17.0.1 20161005

If you need to use different version of intel compilers than the default version in HPC, please check the modules and then load the appropriate one.

module avail intel

output:

----------- /usr/local/share/modulefiles/Core ------------

intel/17 (L)

Both BLAS and LAPACK are installed in the LAPACK module:

module spider lapack

outputs:

------------------------------------------------------

lapack:

------------------------------------------------------

Description:

provides routines for solving systems of simultaneous linear equations, least-squares solutions of linear systems of equations, eigenvalue problems, and singular value problems. The associated matrix factorizations (LU, Cholesky, QR, SVD, Schur, generalized Schur) are also provided, as are related computations such as reordering of the Schur factorizations and estimating condition numbers. Dense and banded matrices are handled, but not general sparse matrices. In all areas, similar functionality is provided for real and complex matrices, in both single and double precision.

Versions:

lapack/3.7.0

------------------------------------------------------

For detailed information about a specific "lapack" module (including how to load the modules) use the module's full name.

For example:

$ module spider lapack/3.7.0

------------------------------------------------------

Load the appropriate module. Do the same for lapack and blas.

module load intel/<version>

Running Lapack/Blas in HPC

Copy the sample LAPACK code (mat_inv.f90) inverts a Non-Hermitian Complex Matrix.

COMPLEX MAT(2,2), WORK(2)

INTEGER PIVOT(2), STAT

MAT(1,1)=(1,1)

MAT(1,2)=(2,-1)

MAT(2,1)=(3,1)

MAT(2,2)=(4,-1)

CALL CGETRF(2,2,MAT,2,PIVOT,STAT)

CALL CGETRI(2,MAT,2,2,WORK,2,STAT)

PRINT *, MAT

END

Interactive

Request a compute node

srun --pty bash

Load the lapack and blas modules

module load lapack

Compile with lapack and blas libraries

ifort mat_inv.f90 -lblas -llapack

Run

./a.out

Batch

Copy the content in your file (let's say job.slurm) in the same directory where the Fortran code is.

#!/bin/bash

#SBATCH --time=00:10:00

#SBATCH -N 1

#SBATCH -n 1

#SBATCH -J lapack-blas-test

#SBATCH -o output

module load lapack

cp -r * $PFSDIR

cd $PFSDIR

ifort mat_inv.f90 -lblas -llapack

./a.out

cp -ru * $SLURM_SUBMIT_DIR

Submit the job

sbatch job.slurm

The output will be in the output file. Verification: type inv([1+i, 2-i; 3+i, 4-i]) in MATLAB:

(-0.6000000,-0.7000001) (9.9999920E-02,0.7000000) (0.4000000,0.3000001)\

(0.1000000,-0.3000000)

Note on GNU compilers

If you want to use the GNU compilers, the above examples work in a similar way. The only differences are:

- load gcc via module load gcc
- replace any instance of ifort by gfortran.