← All resources

Neural Shading Course: Part 1 - Introduction

Auto-discovered

Course Intermediate

NVIDIA Game Developer

Visit resource ↗

Description

Neural Shading Course: Part 1 - Introduction is the opening NVIDIA Game Developer YouTube session of a course on neural and trainable shading inside real-time pipelines.

It frames fidelity limits after hardware growth slowed, defines neural shaders as trainable work that runs inside ordinary shading, compute, or ray-tracing passes, and discusses GPU neural accelerators via cooperative shaders in Vulkan and DirectX. Example problem areas include compression, materials, geometry, radiance caching, and path guiding.

The talk targets graphics engineers who already ship real-time renders. The introduction says prior optimization study is not required, though pipeline familiarity helps.

This part is a free YouTube lecture that opens a longer multi-hour course. Later segments are expected to deepen algorithms and tooling.

It helps you decide where trainable shaders fit before adopting new programming models, and how unused neural hardware flops can serve shading.