Exploration of GPU sharing policies under GEMM workloads

Lately, cloud computing has seen explosive growth, due to the flexibility and scalability it offers. The ever-increasing computational demands, especially from the machine learning domain, have forced cloud operators to enhance their infrastructure with acceleration devices, such as General-Purpose (GP)GPUs or FPGAs. Even though multi-tenancy has been widely examined for conventional CPUs, this is not the case for accelerators. Current solutions support "one accelerator per user" schemes, which can lead to both under-utilization and starvation of available resources. In this work, we analyze the potentials of GPU sharing inside data-center environments. We investigate how several architectural features affect the performance of GPUs under different multi-tenant stressing scenarios. We compare CUDA MPS with the native, default CUDA scheduler and also with Vinetalk, a research framework providing GPU sharing capabilities. Experimental results show that NVIDIA's MPS achieves the best performance in multi-application scenarios, specifically up to X4.5 and X11.2 compared to native CUDA scheduler and Vinetalk respectively.