Microsoft posted a technical brief of the new Hardware-Accelerated GPU Scheduling feature introduced with Windows 10 May 2020 Update, and its latest Windows Display Driver Model (WDDM) version. In a blog post by Steve Pronovost, a tech lead with Microsoft DirectX, Microsoft finally set out to explain what WDDM GPU Scheduling is. Introduced with Windows Vista, WDDM 1.0 introduced GPU scheduling, a software component that allocates workload from multiple sources onto a GPU, prior to which all applications that needed GPU-acceleration would send as much traffic as they could to the GPU driver. With growing complexity in the modern 3D rendering pipeline, the need for a scheduler, not unlike the OS thread scheduler, was needed.

In the following section (which matters), Microsoft went on to detail what Hardware-Accelerated GPU Scheduling is. Apparently, some of the newer generations of GPUs (by NVIDIA, AMD, and Intel), have a hardware component in-built to perform scheduling. When this component is combined with Windows 10 May 2020 Update and a WDDM 2.7-compliant graphics driver, Windows offloads GPU scheduling onto it, freeing up some CPU resources and potentially reducing latencies at various stages of the graphics rendering pipeline. Windows continues to exhibit control over scheduling, but by talking to this scheduling component instead of a CPU-executed software stack."Windows continues to control prioritization and decide which applications have priority among contexts. We offload high frequency tasks to the GPU scheduling processor, handling quanta management and context switching of various GPU engines," writes Pronovost. "The new GPU scheduler is a significant and fundamental change to the driver model. Changing the scheduler is akin to rebuilding the foundation of a house while still living in it. To ensure a smooth transition we are introducing the new scheduler as an early-adopter, opt-in feature. During the transition we will gather large scale performance and reliability data as well as customer feedback."

Microsoft intends for future versions of Windows and DirectX to evolve more in the direction of reducing latencies by offloading the CPU - a full-circle from the company's efforts since Windows XP and Vista to minimize hardware-accelerated signal processing in favor of host-signal processing (the main reason behind the deprecation of DirectSound). "The goal of the first phase of hardware accelerated GPU scheduling is to modernize a fundamental pillar of the graphics subsystem and to set the stage for things to come… but that's going to be a story for a another time," he says.