GDC 2025 Session Viewer

GPU work graphs represent a paradigm shift in GPU programmability by allowing the GPU to self-schedule work without going through the host. It allows for high-performance GPU work creation with fully parallel producer-consumer networks, efficient recursive algorithms, expansion and compaction, GPU-driven draw calls, and many other things.
Join AMD's Max Oberberger for a detailed look at the GPU work graphs API and mesh nodes extension and learn how they are going to enable new forms of GPU-driven applications. Max will show different workloads that can benefit from GPU work graphs today, including an exclusive first look into our latest research. The audience will get an in-depth overview of the current state of the GPU work graphs API and everything they need to get started today, including sample applications and tutorials.

Takeaway

The audience will get a brief introduction to the GPU work graphs API and mesh nodes extension. After this talk, they'll understand the design mentality behind GPU work graphs, the expected performance characteristics, and which workloads are well suited for GPU work graphs. As a result, they'll be able to make informed decisions as to where GPU work graphs can and cannot be applied in their engine and graphics development work.

Intended Audience

The intended audience for this talk is experienced graphics developers familiar with the D3D12/Vulkan and other similar explicit APIs.