OpenMP and MPI Duke it Out in Benchmarks

Douglas Eadlne over at Linux Magazine writes that benchmarking OpenMP vs. MPI is a recipe for surprise:

In a previous column, I had an opportunity to test OpenMP and MPI on a brand new 8-core SMP machine. My interest was to see how well MPI codes worked on an SMP platform. The actual machine was an 8-way Intel server that used two Clovertown processors (4-cores per socket). I also understood that many things other than the programming language (i.e. threads or messages) could effect the result, but I just wanted to get a feel for what would happen.The results were rather interesting (see the column). and there was no clear winner. The OpenMP should “blow away” MPI on an SMP assumption did not hold up.”

Comments

  1. To be honest I’m not surprised, a few years ago a friend of mine benchmarked a crash code which had both SMP and MPI versions on an 8 way Opteron SMP box and found that the MPI version beat the SMP version.

    Our guess at the time was that MPI forced the programmer to think about the parallelism whilst OpenMP let them think that they could just leave it up to the compiler.

    But that’s pure guesswork..

  2. i suggest everyone who want to learn OpenMP to have a glance at

    openmp.blogspot.com

    its a good site to begin learning OpenMP, i found it very helpful…