SwitzerlandManufacturing2023 – ongoingSTMicroelectronics
Reinforcement-learning chip design on Azure
STMicroelectronics reported more than 3x higher design exploration productivity with DSO.ai
Problem
Exploring chip designs to meet power, performance and area requirements required substantial engineering effort. Manual experimentation limited how many alternatives engineers could examine.
Approach
STMicroelectronics used Synopsys DSO.ai on Microsoft Azure alongside Synopsys physical implementation tools. The system used reinforcement learning to search design alternatives. ST completed a commercial tape-out and reported faster implementation of a new Arm core.
Results
Takeaway
ST measured productivity in design exploration. The reported multiplier did not describe the entire chip development cycle.
Categories
- Size
- Large enterprise
- Industry
- Manufacturing
- Business function
- IT and R&D
- Innovation type
- Process innovation
- AI technology
- Optimization
- Adoption
- Vendor solution
- Evidence
- Vendor case study
Sources
Source summaries
Synopsys identified STMicroelectronics as a customer that completed a commercial tape-out using DSO.ai on Azure. An ST design director reported more than 3x higher exploration productivity and said the design exceeded its power, performance and area goals.
Full details are in the original sources.
