Abstract:On the advent of the slow death of Moore's law, the silicon industry is moving towards a new era of chiplets. The automotive industry is experiencing a profound transformation towards software-defined vehicles, fueled by the surging demand for automotive compute chips, expected to reach 20-22 billion by 2030. High-performance compute (HPC) chips become instrumental in meeting the soaring demand for computational power. Various strategies, including centralized electrical and electronic architecture and the innovative Chiplet Systems, are under exploration. The latter, breaking down System-on-Chips (SoCs) into functional units, offers unparalleled customization and integration possibilities. The research accentuates the crucial open Chiplet ecosystem, fostering collaboration and enhancing supply chain resilience. In this paper, we address the unique challenges that arise when attempting to leverage chiplet-based architecture to design a holistic silicon solution for the automotive industry. We propose a throughput-oriented micro-architecture for ADAS and infotainment systems alongside a novel methodology to evaluate chiplet architectures. Further, we develop in-house simulation tools leveraging the gem5 framework to simulate latency and throughput. Finally, we perform an extensive design of thermally-aware chiplet placement and develop a micro-fluids-based cooling design.

What problem does this paper attempt to address?

### Problems Addressed by the Paper This paper aims to explore the application of Chiplet technology in autonomous vehicles to address the challenges faced by the current automotive electronic and electrical architecture (E/E architecture). Specifically, the paper focuses on the following major issues: 1. **The End of Moore's Law**: As Moore's Law gradually becomes obsolete, the silicon industry is entering a new era. Chiplet technology is considered an effective solution to this problem. 2. **Growing Demand for Automotive Computing Chips**: It is expected that by 2030, the market for automotive computing chips will reach $20 billion to $22 billion. High-performance computing (HPC) chips are crucial to meet the ever-increasing computational demands. 3. **Transformation of Automotive Electronic Architecture**: The automotive industry is shifting from hardware-dominated to software-defined vehicles, requiring the automotive electronic architecture to transition from a decentralized, domain-specific model to a centralized, cross-domain model. This transformation brings challenges in complexity and security. 4. **Advantages and Challenges of Chiplet Technology**: Chiplet technology, by decomposing system-on-chip (SoC) into functional modules, offers higher customization and integration possibilities. However, this technology also faces challenges in supply chain, thermal management, and communication protocols. 5. **Open Chiplet Ecosystem**: Establishing an open chiplet ecosystem to promote collaboration among various parties, enhance supply chain resilience, and reduce lock-in effects. ### Main Research Content 1. **Chiplet Microarchitecture Design**: The paper proposes a throughput-oriented microarchitecture suitable for advanced driver-assistance systems (ADAS) and infotainment systems. 2. **Chiplet Architecture Evaluation Method**: A new method is developed to evaluate the performance of chiplet architectures. 3. **Simulation Tool Development**: Internal simulation tools are developed using the gem5 framework to simulate latency and throughput. 4. **Thermal Management and Packaging Design**: A thermally-aware chiplet placement method is proposed, and a microfluidic-based cooling design is developed. ### Application Areas 1. **Advanced Driver-Assistance Systems (ADAS) and Autonomous Driving (AD)**: Chiplet technology can meet the high-performance computing, low latency, and high bandwidth requirements of ADAS and AD systems. 2. **Infotainment Systems**: Chiplet technology can help reduce the cost of ownership, achieve separation of safety and non-safety-related applications, and support virtualization, thereby promoting the development of software-defined digital cockpits. ### Conclusion Through a comprehensive analysis of the cost, performance, and power consumption of chiplet technology, the paper demonstrates the advantages of chiplets in optimizing these key factors. Particularly in the automotive computing stack, chiplet technology has significant application prospects and can address various challenges faced by the current automotive electronic architecture.

Chiplets on Wheels: Review Paper on Holistic Chiplet Solutions for Autonomous Vehicles

Automated Design of Chiplets

Review of Chiplet-Based Design: System Architecture and Interconnection

Furthering Moore’s Law Integration Benefits in the Chiplet Era

ECO-CHIP: Estimation of Carbon Footprint of Chiplet-based Architectures for Sustainable VLSI

RapidChiplet: A Toolchain for Rapid Design Space Exploration of Chiplet Architectures

Modular High-Performance Computing Using Chiplets

The evolution of heterogeneous integration and packaging for the age of chiplets

Performance Implications of Multi-Chiplet Neural Processing Units on Autonomous Driving Perception

Advanced Packaging Design Platform for Chiplets and Heterogeneous Integration

Evaluating Vulnerability of Chiplet-Based Systems to Contactless Probing Techniques

Challenges and Opportunities to Enable Large-Scale Computing via Heterogeneous Chiplets

Architecture of Computing System based on Chiplet

Chiplets and the Codelet Model

The Survey of Chiplet-based Integrated Architecture: An EDA perspective

A fast and simple method to place sub-die under Chip-let architecture based on thermal simulation

Hybrid chips to enable a sustainable internet of things technology: opportunities and challenges

Advanced Packaging Technology Platforms for Chiplets and Heterogeneous Integration

The Decomposition and Combination Paradigms of Chiplet-Based Integrated Chips

BRYT: Data Rich Analytics Based Computer Architecture for A New Paradigm of Chip Design

From 2.5D to 3D Chiplet Systems: Investigation of Thermal Implications with HotSpot 7.0