Chips & computingResearch findings
ThermoDSE · A Thermal-Aware Design Space Exploration Framework for Chiplet-Based DNN Accelerators
Original published 2026.07.08 16:30 KSTAdded to Blackdeck 2026.11.17 10:45 KST
What has changed?
A unified framework jointly considers thermal and area constraints during design space exploration.
Key points
- Chiplet-based DNN accelerators provide scalable performance and yield balance for modern AI workloads.
- ThermoDSE jointly considers architecture, task orchestration, and inter-chiplet communication under thermal and area constraints.
- Achieves up to 3.5x improvement in Energy-Delay-Inverse-Yield over Simba and other baselines.
Connect it to your work
Achieves up to 3.5x improvement in Energy-Delay-Inverse-Yield compared to state-of-the-art methods.
What remains unproven
The abstract reports simulation results; real-world performance is not established by this input.
Originals used for this note
Links open the original publisher in a new tab.