OpenMP Microbenchmarks V2.0

Welcome to the EPCC OpenMP Microbenchmarks V2.0 page.

Overheads due to synchronisation, loop scheduling and array operations are an important factor in determining the performance of shared memory parallel programs. We have designed a set of microbenchmarks to measure these classes of overhead for language constructs in OpenMP. Both Fortran 90 and C versions of the microbenchmarks are available for download.

These codes represent a significant update to version 1.0 of the microbenchmarks released in 1999. The codes are designed to be used with an OpenMP 2.0 compatible compiler. However, as not all users will have access to such a compiler, OpenMP 1.0/2.0 features can be toggled via compiler flags. A link to the major changes/additions made in version 2.0 is provided below.

If required, version 1.0 of the OpenMP Microbenchmarks can still be found here

Comments, suggestions, bug reports and results are all very welcome. Please email Mark Bull or Fiona Reid

Webpage maintained by www@epcc.ed.ac.uk