Dream Chip and Cadence Demo Automotive SoC Featuring Tensilica AI IP at embedded world 2024
Cadence verification and RTL-to-GDS digital full-flow tuned for automotive safety, quality and reliability requirements
June 19, 2024 -- At embedded world 2024, Cadence and Dream Chip demonstrated Dream Chip’s latest automotive SoC, which features the Cadence® Tensilica® Vision P6 DSP IP and Cadence design IP controllers and was taped out using the complete Cadence® Verification solution and full-flow digital implementation, including signoff. The advanced driver-assistance system (ADAS) application software, operating on the Tensilica AI IP within the Dream Chip development board, was a highlight of Cadence's booth during the exhibition held in Nuremberg, Germany, from April 9 to 11.
This innovative SoC utilizes the GlobalFoundries 22FDX® technology platform, which is ready and proven for energy-efficient automotive customers. The SoC is designed in a publicly funded project, ZuSE-KI-Mobil (ZuKIMo). ZuKIMo focuses on developing an AI accelerator system on chip (SoC) for advanced driver assistance systems (ADAS). Manufactured in Dresden, Germany, the SoC supports the local European semiconductor supply chain for automotive applications.
The Cadence Verification, Implementation and Signoff solutions enable automotive designers to quickly deliver first-time silicon success while achieving safety, quality and reliability goals, enabling certification up to ISO 26262 ASIL-D, with minimal impact on system power, performance and area (PPA) targets. This cutting-edge SoC leverages the capabilities of Cadence verification and RTL-to-GDS digital full-flow solutions, meticulously fine-tuned to meet the stringent safety, quality and reliability standards required in the automotive industry.
It also leverages the optimal ADAS features with the industry-leading Tensilica Vision P6 DSP and AI cores by incorporating Tensilica AI processors and a sophisticated array of Digital IP controllers. These include Cadence’s pioneering NPU AI IP, multi-functional Ethernet 10/100/1G MAC, essential control features such as SPI and RTC Control, and SD 6.0/eMMC 5.1 Host Controller and PHY. This strategic integration of hardware is designed to power energy-efficient ADAS applications that utilize evolving technologies like lidar, radar and vision-based systems at the edge, significantly enhancing the automotive-grade quality, safety and reliability of autonomous, connected and electrified (ACE) vehicles.
Commenting on the milestone, Dr. Jens Benndorf, managing director and CEO of Dream Chip Technologies, said, “We collaborated closely with Cadence to provide a full automotive solution highlighting Cadence Tensilica AI-at-the-edge performance along with functional safety and the latest image signal processing to automotive OEMs and Tier-1s. Dream Chip fully implemented and signed-off the energy-efficient automotive SoC from architecture spec to running software using the Cadence Verification solution along with the Cadence Digital EDA full-flow on the GlobalFoundries 22FDX process, showing that Dream Chip is an optimal design house choice for advanced automotive SoCs.”
David Glasco, VP of research and development, Compute Solutions Group at Cadence, added, “The choice of Tensilica IP by the ZuKIMo coalition along with Cadence’s focus on design innovation and excellence across all aspects of the safety-critical and high-reliability IP and the design and verification flow has resulted in this powerful ADAS demo.”
The embedded world demo highlighted the advanced SoC capabilities, demonstrating the power and flexibility of Cadence’s Tensilica IP in driving optimal ADAS features with the industry-leading Vision and AI cores.
See a video of the demo at the Cadence embedded world 2024 booth:
Related Semiconductor IP
Related News
- Mirabilis Design Accelerates SoC Development with New System-Level IP Library for Cadence Tensilica Processors
- Dream Chip Technologies tapes out a 10-TOPS SoC in 22nm with a novel AI Accelerator and an Automotive Functional Safety Processor
- Cadence Expands Tensilica Vision Family with Radar Accelerator and New DSPs Optimized for Automotive Applications
- TSMC and Cadence Collaborate to Deliver AI-Driven Advanced-Node Design Flows, Silicon-Proven IP and 3D-IC Solutions
Latest News
- Will RISC-V reduce auto MCU’s future risk?
- Frontgrade Gaisler Launches New GRAIN Line and Wins SNSA Contract to Commercialize First Energy-Efficient Neuromorphic AI for Space Applications
- Continuous-Variable Quantum Key Distribution (CV-QKD) system demonstration
- Latest intoPIX JPEG XS Codec Powers FOR-A’s FA-1616 for Efficient IP Production at NAB 2025
- VeriSilicon Launches ISP9000: The Next-Generation AI-Embedded ISP for Intelligent Vision Applications