Remapping With the GDE Method
This chapter is devoted to the application of the GDE algorithm in the remapping of data parallel computations. As we have shown in Chapter 6 the inferiority of the diffusion method in static workload model, we will not discuss practical implementations of the diffusion method in this chapter.
KeywordsComputational Fluid Dynamic Unstructured Grid Processor Utilization Computational Graph Grid Partitioning
Unable to display preview. Download preview PDF.