Adaptive Techniques for Dynamic Processor Optimization

1. Front Matter

   Pages i-xvi

   PDF

2. Technology Challenges Motivating Adaptive Techniques

   • David Scott, Alice Wang

   Pages 3-23

3. Technological Boundaries of Voltage and Frequency Scaling for Power Performance Tuning

   • Maurice Meijer, José Pineda de Gyvez

   Pages 25-47

4. Adaptive Circuit Technique for Managing Power Consumption

   • Tadahiro Kuroda, Takayasu Sakurai

   Pages 49-73

5. Dynamic Adaptation Using Body Bias, Supply Voltage, and Frequency

   • James Tschanz

   Pages 75-93

6. Adaptive Supply Voltage Delivery for Ultra-dynamic Voltage Scaled Systems

   • Yogesh K. Ramadass, Joyce Kwong, Naveen Verma, Anantha Chandrakasan

   Pages 95-122

7. Dynamic Voltage Scaling with the XScale Embedded Microprocessor

   • Lawrence T. Clark, Franco Ricci, William E. Brown

   Pages 123-143

8. Sensors for Critical Path Monitoring

   • Alan Drake

   Pages 145-174

9. Architectural Techniques for Adaptive Computing

   • Shidhartha Das, David Roberts, David Blaauw, David Bull, Trevor Mudge

   Pages 175-205

10. Variability-Aware Frequency Scaling in Multi-Clock Processors

    • Sebastian Herbert, Diana Marculescu

    Pages 207-227

11. Temporal Adaptation – Asynchronicity in Processor Design

    • Steve Furber, Jim Garside

    Pages 229-247

12. Dynamic and Adaptive Techniques in SRAM Design

    • John J. Wuu

    Pages 249-272

13. The Challenges of Testing Adaptive Designs

    • Eric Fetzer, Jason Stinson, Brian Cherkauer, Steve Poehlman

    Pages 273-301

14. Back Matter

    Pages 303-304

    PDF

The integrated circuit has evolved tremendously in recent years as Moore’s Law has enabled exponentially more devices and functionality to be packed onto a single piece of silicon. In some ways however, these highly integrated circuits, of which microprocessors are the flagship example, have become victims of their own success. Despite dramatic reductions in the switching energy of the transistors, these reductions have kept pace neither with the increased integration levels nor with the higher switching frequencies. In addition, the atomic dimensions being utilized by these highly integrated processors have given rise to much higher levels of random and systematic variation which undercut the gains from process scaling that would otherwise be realized. So these factors—the increasing impact of variation and the struggle to control power consumption—have given rise to a tremendous amount of innovation in the area of adaptive techniques for dynamic processor optimization. The fundamental premise behind adaptive processor design is the recognition that variations in manufacturing and environment cause a statically configured operating point to be far too inefficient. Inefficient designs waste power and performance and will quickly be surpassed by more adaptive designs, just as it happens in the biological realm. Organisms must adapt to survive, and a similar trend is seen with processors – those that are enabled to adapt to their environment, will be far more competitive.

• Naffziger
• SRAM
• SRAM design
• Sensor
• Wang
• XScale embedded microprocessor
• circuit
• dynamic processor optimization
• microprocessor
• performance tuning
• software
• supply voltage
• testing
• ultra-dynamic voltage scaled systems
• voltage and frequency scaling

• Texas Instruments, Dallas, USA

  Alice Wang

• Advanced Micro Devices, Fort Collins, USA

  Samuel Naffziger

Policies and ethics