DMA for complex System-on-chip in AI, Automotive and IoT
The explosion in data traffic due to the advent of Artificial Intelligence, machine learning, automotive and IoT prompts for new System-on-Chip architectures that allow system developers to build high-performance applications that are able to ingest and process huge amounts of heterogeneous data.
This new breed of System-on-Chip combine an increasing number of processing units, which can be CPU cores, GPU cores, custom accelerator engines, as well as different types of memory subsystems such as RAM cache, DDR and HBM.
To allow these on-chip components to communicate, System-on-Chip need efficient data movers also known as DMA engines that can support hundreds of concurrent, high-performance communication channels.
To address the increasing demand for a scalable high-performance data moving engine, PLDA has developed a DMA controller engine dubbed vDMA, which combines the features and performance expected by designers of next generation System-on-Chip.
Posted on Tuesday Feb. 19, 2019