MosChip Supports EMASS’s RISC-V Edge AI SoC - Embedded Computing Design

Abstract

MosChip has partnered with EMASS to provide development support for EMASS's upcoming RISC-V Edge AI System-on-Chip (SoC). This collaboration focuses on creating specialized, highly efficient processing solutions tailored for demanding Artificial Intelligence applications at the network edge. The project underscores the increasing adoption of the open-source RISC-V architecture in critical, high-performance embedded systems.

Report

Key Highlights

• Collaboration: MosChip is providing critical design and engineering support services to EMASS.
• Product Focus: The partnership centers around the development of EMASS’s specialized System-on-Chip (SoC).
• Core Architecture: The chip utilizes the open-source RISC-V Instruction Set Architecture (ISA).
• Target Application: The SoC is specifically designed for high-performance Artificial Intelligence applications running at the network edge (Edge AI).

Technical Details

• Architecture Standard: The core instruction set relies on RISC-V, which allows EMASS maximum flexibility and customization compared to proprietary ISAs.
• SoC Specialization: Being an "Edge AI SoC" implies the design incorporates specialized hardware accelerators (potentially NPU/DSP blocks) integrated alongside the RISC-V processor cores to maximize AI inference efficiency while minimizing power consumption.
• MosChip's Contribution: MosChip is supporting the complex physical design, integration, and verification processes required to bring a cutting-edge, customized RISC-V chip to tape-out and production.

Implications

• RISC-V Validation: This project further validates RISC-V as a serious contender and a robust foundation for building high-complexity silicon targeting demanding, high-growth markets like Artificial Intelligence.
• Ecosystem Maturation: The involvement of specialized design services companies like MosChip demonstrates the increasing professional support and maturity available within the RISC-V ecosystem, making it easier for new players like EMASS to enter the custom silicon market.
• Edge Computing Advancement: By focusing on Edge AI, the partnership contributes to the development of highly optimized hardware critical for distributing AI processing away from the cloud, leading to faster, more private, and more reliable embedded solutions.