(Credit: Joseph Maldonado/PCMag)
AMD is demoing a new technique that promises to run PC games more efficiently, potentially freeing up resources and speeding up rendering times.
The technique is called the Work Graphs API, which Microsoft released last week. On Monday, its partner on the project, AMD, discussed how developers can use the API to offload more of the game rendering to the GPU, bypassing the need to rely on a PC’s CPU for instructions.
“For many years, game developers have dreamt of a fully GPU-driven renderer where the whole scene processing takes place on the GPU,” AMD architect Matthäus Chajdas says in a blog post.
Currently, a PC game taps the CPU to calculate the GPU workload, which can require constant callbacks to the chip. This can slow down processing times and lead to inefficient resource allocation. The Work Graphs API tries to address these shortcomings by allowing the GPU’s processing threads to schedule the work instructions, resolving the potential CPU bottleneck.
To demo the technology, Chajdas created a 3D environment of a hillside castle town that’s rendered using the Work Graphs API on an AMD Radeon RX 7900 XTX graphics card. The technology is able to efficiently generate the entire environment. At the same time, the API is used for “procedural generation,” by adding more details to the scene, including realistic shadows and the grass on the ground.
In his blog post, Chajdas also noted that Work Graphs can run about 39% faster than Executeindirect, an older API designed to run instructions directly from the GPU.
(Credit: AMD)But one big question is what Work Graphs will mean for games: Might we see a sizable increase in frame rates or faster loading times? While speaking at the Games Developer Conference in San Francisco, Chajdas said it was too early to say how Work Graphs API will translate to gaming performance improvements.
The other issue is that it'll be up to game developers to implement the API, which will likely take time. But to encourage adoption, AMD and Microsoft plan on releasing new updates to the Work Graphs API in the future.