What is Hardware-Accelerated GPU Scheduling Explained

As part of its ongoing effort to optimize the PC gaming experience, Microsoft introduced a powerful but often misunderstood feature in the Windows 10 May 2020 Update (2004): Hardware-Accelerated GPU Scheduling. This feature fundamentally changes how the operating system interacts with the graphics card, offloading a significant portion of memory management from the CPU directly to the GPU’s dedicated scheduling processor. The goal is to reduce latency, improve performance, and create a more responsive system, particularly under heavy graphical loads. This guide delves into what this feature is, the problem it solves, and how it works under the hood.

The Problem: The CPU as a Middleman for VRAM

In the traditional Windows graphics pipeline (using the Windows Display Driver Model, or WDDM), the CPU played a very active role as a high-level scheduler. When a game or application needed to render a frame, the process looked something like this:

  1. The game engine generates rendering commands and data.
  2. These commands are sent to the CPU.
  3. The Windows OS scheduler, running on the CPU, collects these requests from all running applications into batches.
  4. The CPU prioritizes these batches and then submits them to the GPU driver.
  5. Finally, the GPU driver translates these commands and feeds them to the GPU for execution.

This CPU-centric model worked well for years, but it has inherent limitations. The CPU acts as a middleman, introducing a layer of overhead and latency. The constant task of collecting, prioritizing, and scheduling graphics work for the GPU consumes CPU cycles that could be used for other tasks, like game logic or physics. This bottleneck can become particularly noticeable in CPU-bound gaming scenarios, where the CPU is already struggling to keep up with the GPU’s demands.

Introducing Hardware-Accelerated GPU Scheduling: Letting the GPU Take Control

Hardware-Accelerated GPU Scheduling (HAGS) allows the GPU to take direct control over the management and scheduling of its own Video RAM (VRAM). Instead of the CPU’s high-level scheduler dictating every move, the responsibility is handed over to a dedicated, high-priority scheduling processor built directly into modern GPUs. This specialized hardware is much more efficient at managing its own memory than the general-purpose CPU.

With HAGS enabled, the process is streamlined:

  1. The game engine generates rendering commands.
  2. These commands are sent to the GPU driver with minimal processing by the CPU.
  3. The GPU’s dedicated scheduling processor directly manages the VRAM, pulling the data it needs and scheduling the execution of rendering tasks in the most efficient order.

By removing the CPU from this critical path, the system reduces latency between when a frame is requested and when the GPU actually starts working on it. This results in a more efficient and responsive graphics pipeline.

How HAGS Works Internally: A Shift in WDDM

The magic behind HAGS lies in an evolution of the Windows Display Driver Model. The GPU has always had a low-level scheduler to manage its execution units (the CUDA cores or Stream Processors), but the high-level memory management was handled by Windows. HAGS changes this by allowing the GPU to expose its native, hardware-based memory scheduling capabilities directly to the operating system.

The Windows OS still manages the overall priority between different applications (e.g., giving your active game higher priority than a video playing in the background), but the fine-grained, frame-by-frame memory scheduling is now handled by the GPU itself. This is a more efficient division of labor: the OS handles the “what” (which app is important), and the GPU hardware handles the “how” (the most efficient way to render that app’s frames).

This is analogous to a workshop manager (the CPU) who used to tell an expert craftsman (the GPU) the exact order in which to pick up every tool and piece of wood. Now, with HAGS, the manager simply hands the craftsman the blueprints (the rendering commands) and trusts them to manage their own workbench (VRAM) and tools in the most efficient way they know how. The result is the same, but it’s achieved faster and with less micromanagement.

Key Benefits of Hardware-Accelerated GPU Scheduling

  • Reduced Latency: By cutting out the CPU middleman, the time it takes for rendering commands to be executed by the GPU is reduced. This can lead to lower input lag and a more responsive feel in games.
  • Improved Performance (in some scenarios): While HAGS doesn’t magically create more frames, it can provide a performance uplift, particularly in CPU-bound situations. By freeing up the CPU from scheduling tasks, it can devote more cycles to other game-related processing. The performance gains are typically modest, often in the low single-digit percentages, but can help smooth out frame rate dips.
  • Better Video Playback and Multitasking: The benefits aren’t just for gaming. HAGS can also improve the performance of video playback and general system responsiveness when you’re running multiple GPU-intensive applications at once, as the GPU can more intelligently juggle the memory requirements of each task.
  • Foundation for Future Technologies: HAGS is a foundational change that enables other advanced graphics features. For example, it is a key component for realizing the full potential of technologies like Windows DirectStorage, which further optimizes the I/O pipeline by allowing the GPU to directly request assets from a fast NVMe SSD.

Requirements and How to Enable HAGS

To use Hardware-Accelerated GPU Scheduling, you need compatible hardware, an up-to-date operating system, and the latest drivers.

System Requirements:

  • Operating System: Windows 10 (Version 2004 or newer) or Windows 11.
  • GPU:
    • NVIDIA: GeForce 10 series (Pascal) or newer.
    • AMD: Radeon 5600 series (RDNA) or newer.
    • Intel: Xe Graphics (11th Gen Core) or newer.
  • Graphics Driver: A driver that supports WDDM 2.7 or higher. This means you must be on a recent driver from NVIDIA, AMD, or Intel.

For more details on the WDDM specifications, you can refer to the official Microsoft Docs.

How to Turn HAGS On (or Off):

The setting is located within the Windows graphics settings menu.

  1. Right-click on your desktop and select Display settings.
  2. Scroll down and click on Graphics (or “Graphics settings” in older Windows 10 versions).
  3. At the top of the Graphics page, click on Change default graphics settings.
  4. You will see a toggle for Hardware-accelerated GPU scheduling. Turn it on.
  5. You will need to reboot your PC for the change to take effect.

You can follow the same steps to turn the feature off if you encounter any issues.

Potential Drawbacks and Considerations

While HAGS is generally beneficial, there have been some considerations:

  • Early Bugs and Instability: When the feature was first introduced, some users reported instability, crashes, or even performance regressions in certain games with early drivers. These issues have been largely resolved through subsequent Windows updates and driver releases from all major GPU vendors.
  • Variable Performance Gains: The performance impact is not universal. Some games and systems see a noticeable improvement, while others see little to no difference. Its effect is most pronounced on mid-range CPUs that are more likely to be a bottleneck. On very high-end CPUs, the performance gain may be negligible as the CPU has plenty of overhead to handle the scheduling tasks anyway.
  • Software Compatibility: Some software, particularly streaming or recording applications that heavily utilize the GPU encoder (like OBS or NVIDIA ShadowPlay), initially had conflicts with HAGS. It’s always a good idea to ensure all your graphics-intensive applications are up to date.

Frequently Asked Questions

Should I turn Hardware-Accelerated GPU Scheduling on?

For the vast majority of modern gaming PCs that meet the requirements, the answer is yes. The initial bugs have been ironed out, and the feature now provides either a small performance/latency benefit or no change. It’s generally a “set it and forget it” feature that is worth enabling. The only reason to turn it off would be if you are troubleshooting a specific issue with an older game or application that is exhibiting strange behavior.

Will HAGS fix stuttering in games?

It can help in some cases. If your stuttering is caused by a CPU bottleneck where the CPU is failing to feed the GPU in a timely manner, HAGS can alleviate that and smooth out your frame times. However, it will not fix stutter caused by other issues, such as insufficient VRAM, slow storage, or poor in-game frame pacing.

Does HAGS use more GPU power or VRAM?

No. The feature offloads work from the CPU to a specialized, highly efficient scheduling processor on the GPU. This processor is designed for this exact task and its operation has a negligible impact on the GPU’s overall power consumption or thermal output. It also doesn’t change the amount of VRAM a game uses.

Is HAGS the same as Game Mode in Windows?

No, they are different but complementary features. Game Mode is a Windows feature that prioritizes system resources for your game by suspending background activities and updates. HAGS is a lower-level technology that changes how the GPU’s memory is managed. For the best experience, it’s generally recommended to have both Game Mode and HAGS enabled.