Adaptive Power Optimization: Battery-Operated Chip Design
summary
Claims adaptive recalculation using statistical models plus runtime measurements yields concrete improvements and an EDA integration path: a reported 37% power saving, 4% performance gain and 2.8× battery longevity from one device. Method uses toggle‑rate modeling, FSDB/fallback heuristics, checkpoints that emit automated power reports, and power‑net calibration to avoid IR drop and EM. Roadmap includes per‑stage power recomputation (architecture, RTL, post‑route), dashboards that flag missing inputs, and ML to predict activity where vectors are absent.
AI-written summary · about
chapters (8)
- 0:00Introduction and Importance of Battery Life
- 2:00Challenges in Power Calculation
- 4:15Current State and Issues in Power Calculation
- 6:19Adaptive Power Calculation Methodologies
- 8:33Key Milestones in Adaptive Recalculation
- 11:48Automation and Optimization Techniques
- 16:16Implementation Roadmap and Best Practices
- 17:58Conclusion and Key Takeaways
related talks
49mPanel Discussion: Moving Towards...
18mPower Savings & NFV: How to reconcile these 2 constraints ?
01h 01mBuilding the Renewable, Self-Driving Grid
01h 01mContinuous Optimization of MIcroservices using Machine Learning
19mE3: Energy-Efficient Microservices on SmartNIC-Accelerated Servers
17mAutothrottle: A Practical Bi-Level Approach to Resource Management for SLO-Targeted...
29mContinuous tuning of Microservices via Bayesian Optimization
14mPerformance Optimization in High-Volume Transaction Systems