Nvidia, IBM, and university collaborators have a developed an architecture they say will provide fast fine-grain access to large amounts of data storage for GPU-accelerated applications, such as analytics and machine-learning training.
.... more interseting stuff ...
but wait ...
The team plans to open-source details of their hardware and software optimization for others to build such systems. We're reminded of AMD's Radeon Solid State Graphics (SSG) card that placed flash next to a GPU.
Anyone for another EPYC patent court battle :D
Read more ...