GotoBLAS2 uses new algorithms and memory techniques for optimal performance of the BLAS routines. The download site can be found at GotoBLAS2 download
# wget http://cms.tacc.utexas.edu/fileadmin/images/GotoBLAS2-1.13_bsd.tar.gz
# tar -zxvf GotoBLAS2-1.13_bsd.tar.gz
# cd GotoBLAS2
# gmake clean # gmake TARGET=NEHALEM
you will get
GotoBLAS build complete. OS ... Linux Architecture ... x86_64 BINARY ... 64bit C compiler ... GCC (command line : gcc) Fortran compiler ... INTEL (command line : ifort) Library Name ... libgoto2_nehalemp-r1.13.a (Multi threaded; Max num-threads is 8)
you will see the resulting libraries and softlinks
libgoto2.a -> libgoto2_nehalemp-r1.13.a
libgoto2_nehalemp-r1.13.a
libgoto2_nehalemp-r1.13.so
libgoto2.so -> libgoto2_nehalemp-r1.13.so
You can create a /usr/local/GotoBLAS2 and copy the files there and do the PATHING.
If you are having issues, do take a look at Error in Compiling GotoBLAS2 in Westmere Chipsets