Algorithm and Code Development
When moving code from a single processor machine to a parallel computer, both the code and the underlying algorithms must be adapted to the parallel environment. Finding and exploiting the inherent parallelism as well as choosing a suitable parallel programming model (MPI/OpenMP) becomes key issues for porting the code.
Performance tuning and optimization with respect to communication, load balance, synchronization, cache utilization, memory locality, etc. are all important issues in Parallel Algorithm and Code Developement. If you need support or assistance in these matters you are welcome to contact me at support@uppmax.uu.se.
Contact the Application Expert in Algorithm and Code Development Elias Rudberg

