OpenCL & Heterogeneous Computing News
Supporting open standards for parallel programming of heterogeneous systems
Email
* indicates required field
Required fields must be filled in!

News  /  DirectCompute

AMD R-Series APU heterogeneous system architecture for embedded applications

May 24, 2012 by Tony DeYoung

The AMD Embedded R-Series APU combines the new “Piledriver” CPU architecture with discrete-class AMD Radeon HD 7000 Series graphics in a heterogeneous multicore embedded processing platform. It is designed for embedded mid- to high-end graphics-intensive applications such as digital signage, casino gaming, point-of-sale systems and kiosks, as well as parallel-processing-intensive applications spanning medical imaging and security/surveillance.

Software developers can take advantage of APU programmability through OpenCL, OpenGL and DirectCompute.

Traditional embedded solutions using just a CPU run serially with any parallelization occurring in multiprocessor (typically 2-8 cores) systems or virtually via time-splicing control of the individual CPU cores. With the addition of an integrated GPU,  tasks can be distributed over hundreds or thousands of very small and highly-specialized cores enabling parallel execution in a single step to create a parallel compute architecture.

What does GPU compute (i.e. DirectCompute) do for graphics

May 08, 2012 by Tony DeYoung

This video from GDC 2012, explains how GPU compute is used to create realistic lighting effects.

The demo (available for download) uses DirectCompute to cull and manage lights in a scene. This technique also allows for adding one bounce global illumination effects by spawning virtual point light sources where light strikes a surface.The lighting is fully HDR and the material and reflection models take advantage of the GPU to calculate physically accurate light and surface interactions (multiple BRDF equations, realistic use of index of refraction, absorption based on wavelength for metals, etc).

Learn more about GPU Compute for graphics at AFDS 2012 in the Heterogenous Compute and Consumer Graphics track.