The Fortran programming language originated over 50 years ago, and modern variants (e.g. Fortran 95/2003) are the languages of choice for many scientists, engineers and mathematicians for numerical computing.
Members of the University of Sheffield have access to several commercial Fortran compilers all of which have particular strengths and weaknesses.
According to the 2015 Fortran UK compiler comparisons, the NAG compiler was one of the best diagnostic compilers available: https://www.fortran.uk/fortran-compiler-comparisons/intellinux-fortran-compiler-diagnostic-capabilities/
The Windows and Mac versions of the NAG compiler have graphical user interfaces containing full documentation for the Fortran language. Combined with the fact that students can install it on their personal machines, this makes it an ideal compiler for teaching and learning Fortran. Unfortunately, the Linux version is command line only and so is not as easy to use.
On average, programs generated using the Intel Fortran Compiler run faster than those produced by other Fortran compilers (see https://www.fortran.uk/fortran-compiler-comparisons/polyhedron-benchmarks-linux64-on-intel/ for details). At Sheffield, the small number of network licenses mean this compiler is suitable for research use only, i.e. not suitable for teaching use. Our license also allows use of Intel’s debugger and the Math Kernel Library (MKL) which includes optimised versions of the BLAS, LAPACK, FFTW etc.
The PGI Compiler is a useful alternative to the Intel Compiler but its real strength lies when developing code for General Purpose Graphical Processing Units (GP-GPUs). We only have a small number of network licenses so this compiler is not suitable for teaching.
A windows based Fortran and C Compiler suite with excellent debugging facilities. Sheffield has a site license for teaching and research use.
Free to use compiler included in the Linux and Mac OSX operating systems (note: may not be installed by default). gfortran is part of the GNU Compiler Collective (GCC) that includes C/C++ compilers, Windows versions are also available via MinGW or as part of Cygwin. Although performance falls short of the Intel compiler, executable run times are reasonable.
To install in Windows see http://www.mingw.org/
The University of Sheffield has site licenses for the following Fortran numerical libraries:
For queries relating to collaborating with the RSE team on projects: email@example.com
Join our mailing list so as to be notified when we advertise talks and workshops by subscribing to this Google Group.