AMD Introduces Compact Automotive FPGA Package

AMD is expanding its range of automotive FPGAs with a smaller, cheaper version tailored for ADAS sensor applications and in-vehicle infotainment (IVI). The new AMD Artix UltraScale+ XA AU7P 16nm FPGA comes in a smaller package optimized for camera vision and in-vehicle display applications. This chip-scale package is designed to offer increased routing/signal density for I/Os, better solder joint reliability, and enhanced electrical performance.

The FPGA is housed in a compact 9 x 9 mm package, a significant reduction from the previous 19 x 19 mm size. Despite its smaller form factor, the device boasts 81,000 logic cells and 3.8Mbits of memory, making it a powerful solution for automotive applications requiring high performance in a limited space.

With the automotive industry witnessing a growing demand for smaller silicon devices in sensors and digital cockpits, the introduction of the AMD Artix UltraScale+ XA AU7P 16nm FPGA comes at an opportune time. According to Yole Intelligence, the ADAS camera market is expected to grow from $2bn in 2023 to $2.7bn by 2029, highlighting the increasing importance of advanced driver-assistance systems in modern vehicles.

Wayne Lyons, Senior Director of Marketing, Automotive Segment at AMD, emphasized the significance of optimizing form factor, power efficiency, and media processing in the automotive market. Lyons stated, “With the release of this new, small-form-factor Artix UltraScale+ device, AMD continues its commitment to developing solutions that cater to the evolving needs of ADAS and IVI applications."

Customers are already integrating the Artix UltraScale+ AU7P FPGAs into their ADAS edge devices, including thermal and IR cameras, showcasing the versatility and adaptability of these devices in automotive settings. These FPGAs enable automotive designers to handle data ingest and image/video processing for edge sensors, while also facilitating seamless connectivity to in-vehicle displays for enhanced infotainment features.