DI-EDAI
Open-Source Design Tools for the Co-Design of AI Algorithms and AI Chips
DI-EDAI develops open-source design tools for the co-optimized design of AI algorithms and AI chips, directly addressing the tight coupling between software and hardware design in AI applications.
Objectives and Approach
The goal is to develop modern AI chips with a particular focus on security, trustworthiness, and energy efficiency. The project establishes a seamless automated design methodology from AI software algorithms to hardware implementation, alongside an open-source, AI-assisted ecosystem for transparent and sustainable AI system design.
Innovations and Perspectives
Providing low-barrier, high-performance, and secure design tools lowers entry barriers into AI-based chip development, especially for start-ups and SMEs, and strengthens their competitiveness in the rapidly expanding AI technology sector.
Project Coordinator
- Karlsruher Institut für Technologie
Project Partners
- Technische Universität München
- Friedrich-Alexander-Universität Erlangen-Nürnberg
- Rheinland-Pfälzische Technische Universität Kaiserslautern-Landau
This project is funded by the German Federal Ministry for Research, Technology and Space (BMFTR) as part of the Design Initiative Microelectronics.
