LAPACK is big for this system, you should get it direct from: (The .tgz is about 5MB, un zipped 10MB, untarred +10MB more compiled +20MB more, but then keeping only .a library about 5MB.) Using Internet: ftp netlib.org ftp your-email-address cd lapack binary get lapack.tgz quit On your system: mv lapack.tgz lapack.tar.gz gunzip lapack.tar.gz tar -xvf lapack.tar cd BLAS/SRC make cd ../.. cd SRC make If you have not run out of disk space, read the README etc. files Your system may have a large enough /tmp to build liblapack.a You will want a library of the object files in a convenient place so they can be linked in with your applications. You may want to go to TESTING and / or TIMING and make some runs to verify everything is installed OK. Ooopse, if you do not have a Fortran compiler available then use the Internet to get either f2c or g77 free Fortran 77 compilers. f2c can be obtained form netlib.org/f2c get libf77.gz convert Fortran 77 to "C" libi77.gz /src all files g77 can be obtained from prep.ai.mit.edu cd pub cd gnu gcc can be obtained if you do not have a "C" compiler. (Using command line ftp may be necessary, browsers may not provide the required EMail address format for ) Sometimes it is better to go to netlib.bell-labs.com rather than netlib.org One extra step is needed, cd netlib, then do as above. Other software you may need is from /toms These are from collected algorithms of the ACM Transactions On Mathematical Software, TOMS For example 419 is a good root finder for general complex polynomials. Algorithm 535 is a good complex eigenvalue/eigenvector algorithm.