MxGPU: A Case Study on OS-Controlled GPGPU Multiplexing
Marcel Lütke Dreimann und Olaf Spinczyk - Universität Osnabrück

With the growing demand for artificial intelligence and other
data-intensive applications, the demand for graphics processing units
(GPUs) in the server context has also increased. Even though there are
many approaches on multiplexing GPUs, none of the approaches known
to us enable the operating system to coherently integrate GPU resources
alongside CPU resources into a holistic resource management. Due to
the history of GPUs, GPU drivers are still a large, isolated part within
the driver stack of operating systems. This paper aims to conduct a case
study on how a multiplexing solution for GPGPUs could look like, where
the OS is able to define scheduling policies for GPGPU tasks and man-
age GPU memory. We will discuss the architecture of MxGPU, which
offers software-based multiplexing of integrated Intel GPUs. MxGPU
has a tiny code base, which is a precondition for formal verification ap-
proaches and usage in safety-critical environments. Experiments with
our prototype show that MxGPU can grant the operating system con-
trol over GPU resources while allowing more GPU sessions with less
overhead compared to existing work.