Rapidly-growing AI system capabilities are driving new product architectures that demand increased speeds, bandwidths, frequencies and densities. Scalability and configurability are a must. In embedded applications, many AI systems leverage the ease of design and economies of scale of systems-on-modules (SoMs). System-on-modules are an example of make or buy system design. A SoM includes the system compute engine (MCU, FPGA, GPU, etc.) as well as general support circuitry like DRAM, power and … [Read more...]