Concurrency and Interference Analysis of Kernels on GPUs

Heterogeneous systems employing CPUs and GPUs are becoming increasingly popular in large-scale data centers and cloud environments. In these platforms, sharing a GPU across different applications is an important feature to improve hardware utilization and system throughput. However, under scenarios where GPUs are competitively shared, some challenges arise. The decision on the simultaneous execution of different kernels is made by the hardware and depends on the kernels resource requirements. Besides that, it is very difficult to understand all the hardware variables involved in the simultaneous execution decisions, in order to describe a formal allocation method. In this work, we studied the impact that kernel resource requirements have in concurrent execution and used machine learning (ML) techniques to infer the interference caused by the concurrent execution, and to classify the slowdown that results from this interference. The ML techniques were analyzed over the GPU benchmark suites, Rodinia, Parboil and SHOC. Our results showed that, from the features selected in the analysis, the number of blocks per grid, number of threads per block, and number of registers are the resource consumption features that most affect the performance of the concurrent execution.