AMD Radeon datacenter GPUs to power Google's ambitious Stadia gaming streaming service

Google recently announced Stadia, Google's own cloud-based gaming service, where users can play AAA titles on a Chromebook or a smartphone using high-speed internet connection.

AMD has now announced that cloud-optimized Radeon datacenter GPUs will power Google's Stadia cloud gaming service. The GPUs on Google Stadia servers will be capable of handling demanding titles.

What is Google Stadia?

Google Stadia is the search giant's latest venture, which lets users play high-end AAA games on a smartphone, tablet, notebook, Chromebook, and an entry-level laptop. Stadia can offer 4K HDR video game streaming @ 60fps and in the future, Stadia will be able to stream games in true 8K resolution @60fps.

The only requirement for the Stadia is a high-speed internet connection with low latency. Google will use AMD GPU and CPU to render games in a remote location, which will be streamed like a YouTube video.

Google has also announced a new Stadia controller with Wi-Fi connectivity, which connects directly into the server to offer the best possible experience. The controller comes with a dedicated screenshot button and a Google Assistant button for one-touch access, and the overall design looks refreshing compared to a standard gaming controller.

AMD GPU features

Streaming requires a lot of processing capabilities and advanced technologies to tackle unique datacenter challenges from security to scalability. That's why Google has chosen AMD's optimized datacenter GPUs - featuring fast, efficient HBM2 memory and critical datacenter features - for the platform.

AMD GPUs will also support Google by offering software development tools like the AMD Radeon GPU Profiler (RGP), Linux-based, open-source Vulkan driver to help game developers optimize future titles to run on the new GPU-powered platform.

Using RGP, developers can visualize how their application utilizes the GPU and track each rendered frame in real-time, allowing for easy and efficient debugging and optimization through the interoperable RenderDoc debug software.