If you’ve been following the development of your Windows 10 system, the buzzword “hardware-accelerated GPU scheduling” since last year should be no stranger. At the time, the feature was just one option in the setup app, but with the release of drivers, it’s now largely available.
To experience this feature, you need to
Windows 10 May 2020 (20H1/Version 2004) feature update or Build 19041 and later, search for “Winver” in Windows Search to confirm the version
Nvidia GeForce 451.48 or Adrenalin 2020 20.5.1 Beta driver. AMD has now released a beta version, and Intel is planning to update their WDDM 2.7 driver in the near future to support hardware-accelerated GPUs.
After upgrading to the latest version of Windows 10 and drivers, you can turn the feature on/off by setting up an app or registry.
Enabling the feature by setting up the app is as follows
1. Open the Start menu and click on the Set Gear icon.
2. In the Settings app, click System and open the Display tab.
3. In the Multi-Monitor section, select Graphics Settings.
4. Turn on or off the “Hardware Acceleration GPU Scheduling” option.
5. Restart the system.
Enable the application by registry, as follows
1. Open Windows Search
2. Search for “regedit” and select the first option.
3. Open the Registry Editor.
4. In the registry editor, navigate to “HKEY_LOCAL_MACHINE/SYSTEM/CurrentControlSet/Control/GraphicsDrivers”.
5. Find the DWORD option named “HwSchMode”. The value of 2 in this option means that it is enabled, and 1 means that the
6. Restart the system.
7. Turn on settings and display the graphicsetting settings and switch this option.
8. Restart the system.
Hardware-accelerated GPU scheduling is the primary feature of WDDM 2.7, allowing integrated/discrete graphics cards to manage your memory directly, allowing you to achieve better performance, reduce latency, improve average or minimum frame rates, and improve video playback.
With hardware-accelerated GPU scheduling, Windows 10 can now hand control of memory management to a dedicated GPU-based scheduling processor, which in theory should free up the CPU and reduce input lag.
This feature is not obvious on high-end CPUs because they are fully capable of GPU frame processing. “User input is received by the CPU during the ‘N1 frame’ but is not rendered by the GPU until the next frame,” Microsoft said.
However, hardware-accelerated GPU scheduling can prove useful when playing on low-end and mid-range CPUs. According to Microsoft, hardware-accelerated GPU scheduling should also reduce gpu scheduling overhead and modernize a fundamental pillar of the graphics subsystem to “prepare for things of the future.”