The HPC application developer has a variety of tool choices to work from. Essentially the key tools are compilers, libraries, and debuggers.

Commercially developed compiler products include PGI Workstation, the PathScale Compiler Suite, Sun Studio, Absoft Pro Fortran. The prevalent open source compiler is the GNU GCC compiler, the most recent versions of which now support OpenMP, though not auto-parallelization.

Debugging HPC applications necessitates the ability to debug multiple threads and processes, often on multiple machines (meaning MPI must be understood by the debugger). Totalview, PGI's PBDBG, and Allinea DDT are examples of debuggers that provide this capability.

Along with parallel computation with CPUs, it becomes necessary to provide for parallel I/O to deal with ever-increasing dataset sizes. It is critical to understand that for HPC applications in many cases the performance of distributed I/O can become the primary application bottleneck.

The MPI-2 standard includes a standard parallel I/O interface: MPI-IO. These functions are supported on various platforms in the form of a vendor-specific library.