Parallel PHOENICS
Why use PHOENICS parallel
There are two principle reasons why a user may wish to switch to parallel PHOENICS.
They are:
- to obtain computation results in less time, and
- to enable solutions to be made over larger grid domains.
When using the parallel solver the computational domain is shared out over a number processors; each processor performs the computations on their part of the domain in parallel with all the others. Thus a faster the solution can be achieved.
In theory, the more processors used, the faster a solution can be achieved. The following table shows the speed up achieved when carrying out computations on a model with 20 million cells.
It can be seen that there is not a linear relationship between the number of processors and the speed-up ratio. This is because for each additional processor added to the computation pool, there is an additional overhead. Eventually a limit will be reached whereby adding additional processors does not lead to a faster solution time. What this limit is, is dependent on the computational domain under consideration.
What is PHOENICS parallel
The EARTH module uses most of the computing time and therefore it is this part of PHOENICS, which is ported to the parallel computer. The most suitable strategy for parallelising EARTH is grid partitioning (or domain decomposition), where the computational domain is divided into sub-domains. The computational work related to each sub-domain is then assigned to a process. A version of EARTH is replicated over all available processors (or processes) and runs in parallel, exchanging boundary data at the appropriate times.
In the parallel-EARTH implementation, a single process controls the input/output, acting as a server to the other processes; this is usually the ‘Master’ process (PROC-0). The main input operation is to read the required data files from disk, and broadcast them to the other processes. In parallel, each process extracts information specific to the sub-domain it belongs to and continues with the solution procedure.
At the end of computations, the results from each process are sent to the Master process which are re-assembled by the Master, and written on the disk as one PHI file, exactly as in a sequential run (an option for writing individual PHI files also exists). A single result file is also written by the Master process (there is also an option for writing individual result files).
For more information see:
PHOENICS gives you access to the expertise and services of CHAM, the pioneers of
computer simulation of flow and stress.