Dam Break Tutorial

The 2D Dam Break tutorial has been tested, using the 2nd part of the tutorial, namely the multiphase/interFoam/laminar/damBreakFine.

Following the tutorial, but also employing MPICH_PTL_EAGER_LONG, the simulation was run for numberOfSubdomains $={1, 2 ,4, 9 \mbox{and} 16}$, and the resultant times are displayed graphically in figure 2 and presented in table 2.

Figure: Time (secs) for the Dam Break Fine tutorial, where `damBreakFine' is the time to run the tutorial and `damBreakFine.lin` is the line of perfect scaling

Table 2: Timing and performance results for the 2D Dam Break Fine tutorial

Number	damBreakFine
of cores	Time (Perf)
1	174.4 (-)
2	100.5 (1.74)
4	69.1 (1.45)
9	89.0 (0.77)
16	94.9 (0.94)

It can be seen from table 2 that the optimum number of processors is 4; moreover, it is clear from figure 2 that this problem is simply too small to warrant running in production on HECToR. The results are included here for the sake of comparison, as this case is run often as part of the OpenFOAM tutorial.