Performance of a Code Migration for the Simulation of Supersonic Ejector Flow to SMP, MIC, and GPU Using OpenMP, OpenMP+LEO, and OpenACC Directives

Figure 7

Design in OpenMP for the MIC architecture. (a) For each time iteration, one parallel region is opened and unloaded; (b) the parallel region is maintained until it is necessary to write data to the permanent memory; at that point control is returned to the CPU followed by a reentry to the MIC.

(a)

(b)