Cooperative vectors

NPUs, tensor cores, and matrix math machines will be exposed.

March 20, 2025

Jon Peddie

Matrix and vector operations benefit from cooperative thread behavior for optimized performance. APIs exposing this hardware acceleration were initially termed Cooperative/Wave but are now grouped under linear algebra, reflecting their functionality rather than implementation specifics. Nvidia has demonstrated a cooperative vectors concept, utilizing a GPU-based neural network to approximate complex calculations, potentially enhancing real-time rendering performance. Microsoft plans a CoopVec API preview, with a full release expected later. While immediate visual improvements may be minor, the utilization of these AI engines through tools like segmented shaders is projected to lead to performance gains over time.In January, Microsoft formally introduced its

...

Enjoy full access with a TechWatch subscription!

TechWatch is the front line of JPR information gathering service, comprising current stories of interest to the graphics industry spanning the core areas of graphics hardware and software, workstations, gaming, and design.

A subscription to TechWatch includes 4 hours of consulting time to be used over the course of the subscription.

Already a subscriber? Login below

Cooperative vectors

Enjoy full access with a TechWatch subscription!

This content is restricted

Login