Process Detector (For DVFS and monitoring process variation), TSMC N3E
Deep Vision Adopts SiFive RISC-V to Add OpenCV-Enabled AI Support
Deep Vision licenses SiFive Intelligence X280 processor to deliver greater flexibility and AI inerence pre-processing for many markets
SAN MATEO, Calif., January 17, 2022 – SiFive, Inc., the founder and leader of RISC-V computing, today announced that Deep Vision will integrate SiFive RISC-V processor IP into its next-generation inference accelerators to enable more comprehensive computer vision and voice in edge devices. Specifically, Deep Vision will license SiFive Intelligence™ X280 and SiFive Essential™ S7 processor IP to enhance the flexibility and functionality of their products to better support customers building applications for markets like smart city, smart retail, automotive, and industrial.
SemiCo Research(1) estimates the AI SoC market to reach 25B units by 2027, with revenue reaching $291B, with RISC-V-based AI SoCs unit CAGR of 73.6% by 2027. The SiFive Intelligence X280 processor enables Deep Vision accelerators to offer broader neural network model support and optimizations for floating-point operations. SiFive Intelligence processors feature SiFive Intelligence Extensions to support TensorFlow Lite and a broad range of AI/ML data types, including BFLOAT16. The flexibility and programmability of the SiFive Intelligence X280 processor as an application class AI processor extends the proven capabilities and performance of Deep Vision accelerators to rapidly support new and evolving AI inference models in hardware, as well as AI pre-processing tasks such as image scaling, color conversion, and white noise subtraction. The SiFive Essential S7 processor features real-time, deterministic processing capabilities to enable command and control applications within a heterogeneous compute cluster, alongside other SiFive RISC-V processors. Ideally suited to latency-sensitive edge compute applications, the SiFive Essential S7 processor architecture features enhanced security and real-time determinism while offering area-efficient performance-per-watt.
“We are excited to increase the flexibility and capabilities of our next-generation AI accelerators with the performance and features of SiFive RISC-V processors,” said Ravi Annavajjhala, CEO, Deep Vision. “We can address a broader range of AI use cases and applications, thanks to SiFive Intelligence and SiFive Essential processors, as well as accelerate customer adoption of our solutions. With the AI market evolving rapidly, flexibility and more comprehensive AI support are crucial for success.”
“Combining the strengths of SiFive Intelligence RISC-V processors with Deep Vision inference accelerators is a natural way to quickly and efficiently build new hardware for the rapidly evolving AI processor market,” said Dr. Yunsup Lee, CTO and co-founder, SiFive. “The requirements for modern AI are well matched to RISC-V, RISC-V Vectors, and SiFive Intelligence Extensions, as implemented in the SiFive Intelligence X280. With our industry-leading RISC-V processor IP portfolio, SiFive is perfectly positioned to enable AI companies who seek to quickly develop market-focused platforms.”
Recently, SiFive introduced the SiFive 21G3 update with enhancements and performance improvements for a wide range of products, including support for the recently ratified RISC-V Vector and Hypervisor specifications. For more information, please read our blog covering recent announcements, here.
About Deep Vision
Deep Vision, an Edge AI processor company, develops specialized deep learning solutions for the edge. Deep Vision silicon and software deliver the power, performance, and ease of use application developers need to enable pervasive edge AI. Deep Vision's products are based on groundbreaking research into the architecture and tools needed to build and offer a leading-edge AI processor for a rapidly growing Edge AI market. Application markets include smart retail, driver-monitoring systems, smart city, drones, factory automation, and more. More information is available at DeepVision.io.
As the pioneers who introduced RISC-V to the world, SiFive is transforming the future of compute by bringing the limitless potential of RISC-V to the highest performance and most data-intensive applications in the world. SiFive’s unrivaled compute platforms have enabled leading technology companies around the world to innovate, optimize and deliver the most advanced solutions of tomorrow across every market segment of chip design, including artificial intelligence, machine learning, automotive, data center, mobile, and consumer. With SiFive, the future of RISC-V has no limits. For more information, please visit www.sifive.com.
Search Silicon IP
- Arteris and SiFive Partner to Accelerate RISC-V SoC Design of Edge AI Applications
- BSC, Codeplay and SiFive help accelerate applications on RISC-V thanks to V-extension support in LLVM
- SiFive and ArchiTek Enable Secure, Private, Flexible Edge AI Computing With AiOnIc Processor
- Faraday Unveils RISC-V ASIC Solution to Support Edge AI and IoT SoCs
- DinoplusAI Partners with SiFive to Develop Mission-Critical AI Processor Platform for High Performance Processing with Ultra-Low Latency
- 2026 All-Time High in Store for Global 300mm Semiconductor Fab Capacity After 2023 Slowdown, SEMI Reports
- Designing MPUs/MCUs with Functional Safety
- Sensor Fusion Explores AI to Prep for ADAS, AV Designs
- MIPI A-PHY Specification Levels Up In-Vehicle Connectivity
- Bluetooth Dual Mode V5.3 RF Transceiver IP Cores in 22nm ULP is Available for Immediate Licensing for Smart Audio Chipsets
- OPENEDGES Completes the Tapeout of the 7nm HBM3 Memory Subsystem (PHY & Memory Controller) Test chip
- Verisity Announces Specman Elite Version 4
- BrainChip Begins Taking Orders of Akida AI Processor Development Kits
- Renesas to Acquire Panthronics to Extend Connectivity Portfolio with Near-Field Communication Technology
- SoftBank and EdgeCortix Partner to Jointly Realize Low-latency and Highly Energy-efficient 5G Wireless Accelerators
|E-mail This Article||Printer-Friendly Page|