Installing LAPACK and BLAS

See Linear Algebra software for more references.

First see if these are already installed. You could do this by going to the directory $CLASSHG/codes/lapack/random and typing:

$ make randomsys3.exe

if you get a message like:

ld: library not found for -lblas

then you’ll need to install these.

This was discussed in Lecture 17.

To install BLAS and LAPACK, see the links below and also:

• http://gcc.gnu.org/wiki/GfortranBuild

On some linux systems, including the VM for the class, you can install both BLAS and LAPACK via:

$ sudo apt-get install liblapack-dev

The BLAS can be installed by downloading from http://www.netlib.org/blas/blas.tgz

Put this in desired location, e.g. $CLASSHG/codes/lapack/blas.tgz and then:

$ cd $CLASSHG/codes/lapack
$ tar -zxf blas.tgz    # creates BLAS subdirectory
$ cd BLAS
$ gfortran -O3 -c *.f
$ ar cr libblas.a *.o  # creates libblas.a

To use this library:

$ gfortran -lblas -L$CLASSHG/codes/lapack/BLAS  \
           program.f90

A small subset of LAPACK routines are in $CLASSHG/codes/lapack/lapack-subset. In this directory do:

$ make lib

which should create liblapack.a.

To use this library:

$ gfortran -lblas -L$CLASSHG/codes/lapack/BLAS  \
           -llapack -L$CLASSHG/codes/lapack/lapack-subset \
           program.f90

With these you should be able to run the codes in $CLASSHG/codes/lapack/random, though you may need to add -L flags in the Makefile if you’ve put libraries in nonstandard places as in the above examples.

To get additional LAPACK routines separately, see http://www.netlib.org/lapack/double/ and click on some routine name plus dependencies to get a zip or tar file of the routine and any other routines it calls.

Some references:

• LAPACK
• LAPACK User’s Guide
• blas
• Automatically Tuned Linear Algebra Software (ATLAS)
• http://en.wikipedia.org/wiki/LAPACK