AMD launches second-generation Versal series devices to provide end-to-end acceleration for AI-driven embedded systems
— The first devices in the second-generation Versal series product portfolio use the next-generation AI engine to increase TOPS per watt by up to 3 times, while increasing CPU-based scalar computing power by up to 10 times compared to the first generation —
Nuremberg, Germany (International Embedded Show), April 9, 2024 - AMD (NASDAQ: AMD) today announced the expansion of its AMD Versal™ adaptive system-on-chip (SoC) product portfolio, launching a new The second-generation Versal AI Edge series and the second-generation Versal Prime series adaptive SoC integrate pre-processing, AI inference and post-processing into a single device, providing end-to-end acceleration for AI-driven embedded systems.
These first-generation devices in the second-generation Versal series portfolio build on the first-generation and feature a powerful new AI engine that is expected to deliver up to 3x more TOPS per watt than the first-generation Versal AI Edge series devices , 1 while also delivering a new high Performance - integrated Arm® CPUs are expected to deliver up to 10x the scalar computing power of first-generation Versal AI Edge and Versal Prime series devices2 .
Salil Raje, senior vice president and general manager of AMD's Adaptive and Embedded Computing Group, said : "Demand for artificial intelligence-enabled embedded applications is exploding, driving concerns about the power and footprint of embedded systems. The demand for a single-chip solution that achieves the most efficient end-to-end acceleration within area constraints, relying on more than 40 years of leadership in adaptive computing, these latest generation Versal devices integrate multiple computing engines on a single architecture to provide high computing. Efficiency and performance and scalability from low-end to high-end.”
The second-generation Versal family of devices balances performance, power consumption, footprint, and advanced functional safety and information security. It provides new functions and features to support high-end devices designed for the automotive, industrial, vision, medical, broadcast and professional audio and video markets. Performance edge optimized products .
Powering Subaru’s next-generation ADAS vision system
Subaru Corporation has selected the second-generation Versal AI Edge series for the company's next-generation advanced driver assistance system (ADAS) vision system, called EyeSight . The EyeSight system is integrated into select Subaru models for advanced safety features including adaptive cruise control, lane-keep assist, and pre-collision braking. Subaru is using AMD Adaptive SoC technology in current EyeSight-equipped vehicles .
"Subaru has selected the second-generation Versal AI Edge series to provide next-generation in-vehicle AI performance and safety for future EyeSight-equipped vehicles, " said Satoshi Katahira, general manager of the Advanced Integrated Systems Division and ADAS Development Division , Subaru Corporate Engineering. "The second-generation Versal AI Edge devices are designed to deliver AI inference performance, ultra-low latency and functional safety, bringing cutting-edge AI-based safety features to drivers.”
Second generation Versal AI Edge series
In order to meet the complex processing needs of real-world systems, AMD's second-generation Versal AI Edge series devices use an optimal combination of processors to accelerate all three stages of AI-driven embedded systems :
-
Pre-processing : FPGA programmable logic architecture for real-time processing, superior flexibility to connect various sensors and enable high-throughput, low-latency data processing workflows;
-
AI inference : The vector processor array constitutes the next generation AI engine, which can achieve efficient AI inference;
-
Post-processing : Arm CPU cores provide the post-processing capabilities required for complex decision-making and control in safety-critical applications.
This single-chip intelligence eliminates the need to build multi-chip processing solutions, resulting in smaller, more efficient embedded AI systems and the potential for faster time to market.
Second generation Versal Prime series
AMD's second-generation Versal Prime series combines programmable logic for sensor processing with high-performance embedded Arm CPUs to provide end-to-end acceleration for traditional non-AI embedded systems. Designed to deliver up to 10x more scalar computing power than the original generation , these devices can efficiently perform sensor processing and complex scalar workloads.
With new hard IP for high-throughput video processing, including multi-channel workflows up to 8K, second-generation Versal Prime devices are ideally suited for applications such as ultra-high definition (UHD) video streaming and recording, industrial PCs and more.
Broad and scalable product portfolio
The second-generation Versal AI Edge Series and second-generation Versal Prime Series product combination provides scalability from edge sensors to central computing for AI-driven systems. It features a series of devices with increasing AI and adaptive computing power, allowing customers to choose performance, power consumption and footprint to efficiently achieve application performance and security goals .
Simplify the design cycle
AMD Vivado™ Design Suite tools and libraries help increase productivity and streamline design cycles for embedded hardware system developers, resulting in faster compile times and higher quality results. For embedded software developers, the AMD Vitis™ unified software platform enables embedded software, signal processing and AI design development at the user's preferred level of abstraction, without requiring FPGA experience.
Designers can start with the now-available early access documentation for AMD’s second-generation Versal AI Edge series and second-generation Versal Prime series, as well as first-generation Versal evaluation kits and design tools. AMD expects to provide second-generation Versal series chip samples in the first half of 2025, followed by evaluation kits and system modules (SOM) samples in mid-2025, and is expected to provide mass production chips by the end of 2025.
Swipe to view↓