OpenMP, MPI and CUDA are open-source, cross-platform packages, so in
theory you could work on your own machine at home in some cases.
However, this is a course in parallel programming, so you need a
parallel machine:
- You probably have a dual-core machine at home. You'll be able
to do some of your OpenMP/threads work on it, but in some cases
we'll be using a quad-core machine on CSIF.
- CUDA requires a high-end NVIDIA video card, as we have on some
CSIF machines. If you have this on your machine, you can work there.
- MPI requires a network of machines, such as CSIF, or again, a
single machine with many cores. In principle it could be used on a
single-core machine for debugging purposes.
Norm Matloff
2016-03-30
