AMD partners with Arm to unveil an AI chip family that does preprocessing, inference, and postprocessing on a single silicon – but you’ll have to wait more than 12 months to get actual products
AMD is introducing two new adaptive SoCs: Versal AI Edge Series Gen 2 for AI-powered embedded systems, and Versal Prime Series Gen 2 for classic embedded systems.
Multi-chip solutions typically incur significant overhead, but the single-hardware architecture is not fully optimized for all three AI phases: preprocessing, AI inference, and postprocessing.
To address these challenges, AMD has developed a single-chip heterogeneous processing solution that streamlines these processes and maximizes performance.
Still early days
The Versal AI Edge Series Gen 2 adaptive SoCs provide end-to-end acceleration for AI-powered embedded systems, which the tech giant says is built on a foundation of enhanced safety and security. AMD has integrated a powerful processing system, featuring Arm CPUs and next-generation AI engines, with top-of-the-line programmable logic, creating a device that expertly handles all three phases of computation required in embedded AI applications.
AMD says the Versal AI Edge Series Gen 2 SoCs are suitable for a broad spectrum of embedded markets, including those with high security, high reliability, long lifecycle and safety-critical requirements. Purposes include autonomous driving, industrial PCs, autonomous robots, edge AI boxes and ultrasound, endoscopy and 3D imaging in healthcare.
The integrated CPUs’ processing system includes up to 8x Arm Cortex-A78AE application processors, up to 10x Arm Cortex-R52 real-time processors, and support for USB 3.2, DisplayPort 1.4, 10G Ethernet, PCIe Gen5, and more.
The devices meet the operational requirements of ASIL D/SIL 3 and comply with a range of other safety and security standards. They reportedly offer up to three times the TOPS/watt for AI inference and up to ten times the scalar computing power with powerful CPUs for post-processing.
Salil Raje, senior vice president of AMD’s Adaptive and Embedded Computing Group, said: “Demand for AI-enabled embedded applications is exploding, driving the need for solutions that bring together multiple compute engines on a single chip for the most efficient end- to-end applications. -end acceleration within the power and area limitations of embedded systems. Backed by more than 40 years of leadership in adaptive computing in high security, high reliability, long lifecycle, and safety-critical applications, this latest generation of Versal devices deliver high compute efficiency and performance on a single architecture that scales from the low -end to high-end.”
Early access documentation and evaluation kits for the devices are now available. The first silicon samples of the Versal Series Gen 2 are expected early next year, with production expected to begin in late 2025.