Microsoft experiments with high-end machine learning on Xbox Series X

Microsoft is hiring new software engineers to help create new machine learning algorithms that will push next-gen game on Xbox Seires X/S consoles.

VIEW GALLERY – 2 IMAGES

Microsoft is doing fairly advanced work with Xbox hardware, namely with the software stack that optimizes next-level gaming performance like 120FPS, ray tracing, and 4K 60FPS. The system is powered by its new DirectX 12 suite of software APIs, which includes DirectStorage, Sampler Feedback Streaming, and DirectML, a feature that allows machine learning support on Xbox consoles. DirectML is a game-changer that helps run ultra-demanding games like Flight Simulator on comparatively lower-end console APUs.

The company is currently hiring to push DirectX 12’s hardware-accelerated machine learning to new heights and continue balancing performance targets with ML algorithms.

“Engineering roles are available to direct the future of machine learning and computer graphics technology on Xbox consoles. These roles would focus on researching and implementing machine learning algorithms into software to make traditional rendering techniques more efficient and provide new alternative solutions,” reads the recruitment page.

Microsoft has two job openings for ML work:

DirectML – Xbox Series X and Xbox Series S support Machine Learning for games with DirectML, a component of DirectX. DirectML leverages unprecedented hardware performance in a console, with Xbox Series X benefiting from over 24 TFLOPS of 16-bit float performance and over 97 TOPS (trillion operations per second) of 4-bit integer performance on Xbox Series X. Machine Learning can improve a wide range of areas, such as making NPCs much smarter, providing vastly more lifelike animation, and greatly improving visual quality.

DirectML also helps power Super Resolution upscaling to make in-game visuals sharper and more clear.

According to RedGamingTech, both the Series X and Series S support 4- and 8-bit operations for inference machine learning.

There’s two parts to AI-based machine learning: training and inference. Training is done by supercomputers. Developers feed code into huge datacenters and the AI trains the data. Once the data is ready, it can be distributed and ready via inference.

Combined with other DX12 tools like Sampler Feedback Streaming, which allows devs to efficiently process assets without wasting GPU power, and Variable Rate Shading, which imperceptibly lowers resolution of far-away targets for better frame rates, DirectML poses major advantages for gaming on Xbox consoles, Windows PCs, and Project xCloud.