The industry is divided over programming models for graphics cores embedded in SoCs. Apple was an early advocate of the OpenCL approach, Microsoft uses DirectCompute, Google promotes Renderware and Nvidia is using its proprietary Cuda environment.
A handful of chip designers including ARM, Imagination, MediaTek and Qualcomm are part of the HSA group. However a senior Qualcomm architect said recently the company remains agnostic about its approach awaiting a broader industry consensus.
For its part, Microsoft has taken part in HSA events, but not joined the group.
AMD said the HSA group will provide support for its approach in C++, Java and Python. AMD will host a developer’s summit in San Jose, Calif., Nov. 11-14, likely the venue where the first HSA specs will be formally released.
Click on image to enlarge.
The HSA approach can eliminate some data-copy operations, boosting performance and power efficiency.
The obvious win here is in providing a more usable programming model for CPU/GPU utilization. But It's also worth pointing out that purely CPU applications will benefit, since AMD is basically adding a new, higher-bandwidth memory system that's not specific to the GPU. Yes, your Excel will be able to update quicker...
Join our online Radio Show on Friday 11th July starting at 2:00pm Eastern, when EETimes editor of all things fun and interesting, Max Maxfield, and embedded systems expert, Jack Ganssle, will debate as to just what is, and is not, and embedded system.