Architecture, Simulation and Software Stack to Support Post-CMOS Accelerators: The ARCHYTAS Project

2025-10-18

AI Summary

To address post-CMOS bottlenecks—including high power consumption, low energy efficiency, and poor scalability—in defense AI applications (e.g., autonomous platforms, surveillance systems), this work proposes a unified hardware-software co-design framework supporting photonic computing, in-memory computing (with both volatile and non-volatile memory technologies), and neuromorphic architectures. We innovatively design a system architecture capable of integrating heterogeneous accelerators across multiple paradigms and develop a lightweight, domain-specific software stack. Furthermore, we build a high-fidelity full-system simulation platform to enable early-stage prototyping and validation. Our approach bridges the gap between emerging devices and practical deployment in complex defense AI tasks, establishing a verifiable architectural foundation and toolchain for scalable, energy-efficient intelligent computing in defense systems. (138 words)

Technology Category

• Machine Learning: Hardware-aware ML
• Cognitive Modeling & Cognitive Systems: (Computational) Cognitive Architectures
• Computer Vision: Adversarial Attacks & Robustness

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Security and privacy of machine learning and AI applications
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

ARCHYTAS aims to design and evaluate non-conventional hardware accelerators, in particular, optoelectronic, volatile and non-volatile processing-in-memory, and neuromorphic, to tackle the power, efficiency, and scalability bottlenecks of AI with an emphasis on defense use cases (e.g., autonomous vehicles, surveillance drones, maritime and space platforms). In this paper, we present the system architecture and software stack that ARCHYTAS will develop to integrate and support those accelerators, as well as the simulation software needed for early prototyping of the full system and its components.

Problem

• Designing non-conventional hardware accelerators to overcome AI bottlenecks
• Developing system architecture and software stack for accelerator integration
• Creating simulation tools for early prototyping of full systems

Innovation

• Designs non-conventional hardware accelerators for AI
• Develops system architecture and software stack
• Uses simulation software for early prototyping