Marvell shows AI SSD controller architecture

August 10, 2018 // By Ally Winning
Marvell will use the Flash Memory Summit to demonstrate how it will incorporate NVIDIA’s Deep Learning Accelerator (NVDLA) to provide AI capabilities to a large number of solutions.

The proof-of-concept architecture solution will show how machine learning can accelerate applications with minimal network bandwidth and no host CPU processing. The architecture is anticipated to enable a new age of storage SSD solutions.

By adding NVDLA to its SSD controllers, Marvell will improve efficiencies, reduce power consumption, maximise scalability and optimise the distribution of resources. The scalability of this solution will also engineers to add more offerings and capabilities to their product portfolios. This programmable architecture will enable AI models to be quickly updated to address new use cases.

More information

www.marvell.com

Related news

First NVMe-oF SSD converter controller

First 96-Layer 3D Flash memory SSD

NVMe SSD Controller family for 96-layer SSDs

Host bus adapter support Dual/Quad M.2 SSDs


Vous êtes certain ?

Si vous désactivez les cookies, vous ne pouvez plus naviguer sur le site.

Vous allez être rediriger vers Google.