After last week’s release of NVIDIA’s first hardware accelerated GPU-enabled video card driver, AMD has taken the lead this week to do the same. The Radeon Software Adrenalin 2020 Edition 20.5.1 Beta driver with graphics hardware programming driver (version 20.10.17.04) has been released on the AMD website and, as the name implies, the driver offers support for new programming GPU hardware accelerated Windows 10 technology.
As a quick update, hardware acceleration for GPU programming has been added to the Windows display driver stack with WDDM 2.7 (included in Win10 2004). And, as the name alludes to, it allows GPUs to more directly manage their VRAM. Traditionally, Windows itself has done much of the VRAM management for GPUs, making this a distinctive change in business.
Microsoft has been treating the feature as a relatively unobtrusive development, relative to DirectX 12 Ultimate, they haven’t said much about it, while AMD’s release notes make vague performance-enhancing claims, claiming “By shifting programming responsibilities From software to hardware, this feature has the potential to enhance the responsiveness of the GPU and enable additional innovation in GPU workload management in the future. ”As was the case with the launch of NVIDIA the Last week, don’t expect anything too significant here, otherwise AMD would be promoting performance gains much more, but it’s something to look out for in the long run.
Meanwhile, AMD appears to be taking a cautious approach here. The beta driver has been released outside of its normal launch channels and is only compatible with products that use AMD’s Navi 10 GPUs, making the Radeon 5700, 5600 series and their mobile variants. Support for the Navi 14-based 5500 series is noticeably absent, as is Vega’s support for discrete, integrated GPUs.
Additional details on the driver version, as well as download instructions, can be found on the AMD website in the driver release notes.
Finally, on a tangential note, I’m trying to sit down with The Powers That Be for the next week or so to dig deeper into hardware accelerated GPU programming. Since this is primarily a hardware developer-centric feature, Microsoft hasn’t discussed that much in the consumer or media context. So I’ll dig deeper into the theory behind this: what it’s meant to do, the prospects for future features, and the rationale for presenting it now rather than before (or later). Be sure to come back next week for that.
