Alternative Computing Technologies

1. Alternative Computing Technologies CS 8803 ACT Spring 2014 Hadi Esmaeilzadeh hadi@cc.gatech.edu Georgia Institute of Technology

2. Hadi EsmaeilzadehFrom Khoy, Iran

3. PhD in CSE, University of WashingtonDoug Burger and Luis Ceze 2013 William Chan Memorial Dissertation Award MSc in CS, The University of Texas at Austin MSc and BSc in ECE, University of Tehran

4. Research: ACT LabAlternative Computing Technologies • General-purpose approximate computing • Bridging neuromorphic and von Neumann models of computing • Analog computing • System design for online machine learning • System design for perpetual devices

5. Agenda • Who is Hadi • Course organization • Why alternative computing technologies • How we became and industry of new possibilities • Why we might become an industry of replacement • Possible alternative computing technologies • Quiz # 1

6. Objective • Explore cutting-edge research on new and alternative paradigms of computing • Empower you with higher order critical thinking • Improve your technical writing and speaking • Innovate in alternative computing technologies

7. Format • Seminar course • Reading papers • Critiquing and discussing the papers • Brainstorming about new ideas • Developing new technologies • Mostly your presentations • I will only lecture three times

8. Grading rubric

9. Class presentation • Objective: Communicate and analyze ideas • 4 points: Clearly presenting the key ideas • 1 points: Clear, well-organized slides • 5 points: Stimulating interesting discussion • 1 point bonus

10. Class participation • You have to say somethinginteresting! • By 9pm the night before, two comments/questions • Your new ideas • Critical questions about methodologies and conclusions • Why will the paper be cite • What you learned • Main insights from the papers

11. Critiques • Objective: developing high-order critical thinking • Summary (quarter a page) • Strengths (1-3 sentences) • Weaknesses (1-3 sentences) • Analysis I (1 paragraph) • Analysis II (1 paragraph) • Please read the “The task of the referee by Alan Jay Smith”

12. Reading material for writing critiques Style: The Basics of Clarity and GraceJoseph M. Williams The task of the referee Allen Jay Smith

13. Final project • Groups of two • Options • Implementing a new idea • Extending an existing paper • Re-implement a paper • Survey at least ten papers • Evaluation • Implementation • Writing • Oral presentation

14. Prerequisites • Understand a subset of • VLSI Circuits • Computer architecture • Programming Languages • Machine learning • Do • Programming

15. Agenda • Who Hadi is • Course organization • Why alternative computing technologies • How we became and industry of new possibilities • Why we might become and industry of replacement • Possible alternative computing technologies • Quiz # 1

16. What has made computing pervasive? What is the backbone of computing industry?

17. Networking Programmability

18. What makes computers programmable?

19. von Neumann architectureGeneral-purpose processors • Components • Memory (RAM) • Central processing unit (CPU) • Control unit • Arithmetic logic unit (ALU) • Input/output system • Memory stores program and data • Program instructions execute sequentially

20. Programmability versus Efficiency

21. Programmability versus Efficiency General-Purpose Processors SIMD Units Programmability GPUs FPGAs ASICs Efficiency

22. What is the difference between the computing industry and the paper towel industry?

23. Industry of replacement 1971 2013 ? Industry of new possibilities

24. Can we continue being an industry of new possibilities? Personalizedhealthcare Virtualreality Real-timetranslators

25. Agenda • Who Hadi is • Course organization • Why alternative computing technologies • How we became and industry of new possibilities • Why we might become and industry of replacement • Possible alternative computing technologies • Quiz # 1

26. Moore’s LawOr, how we became an industry of new possibilities Every 2 Years • Double the number of transistors • Build higher performancegeneral-purpose processors • Make the transistors available to masses • Increase performance (1.8×↑) • Lower the cost of computing (1.8×↓)

27. What is the catch?Powering the transistors without melting the chip Moore’s Law W W

28. Dennard scaling: Doubling the transistors; scale their power down Transistor: 2D Voltage-Controlled Switch Dimensions Voltage ×0.7 DopingConcentrations 0.5×↓ 0.7×↓ Capacitance Area 0.5×↓ Power 1.4×↑ Frequency Power = Capacitance × Frequency × Voltage2

29. Dennard scaling broke: Double the transistors; still scale their power down Transistor: 2D Voltage-Controlled Switch Dimensions Voltage ×0.7 DopingConcentrations 0.5×↓ 0.7×↓ Capacitance Area 0.5×↓ Power 1.4×↑ Frequency Power = Capacitance × Frequency × Voltage2

30. Dark siliconIf you cannot power them, why bother making them? 0.5×↓ Area Dark Silicon 0.5×↓ Power Fraction of transistors that need to bepowered off at all timesdue to power constraints

31. Looking backEvolution of processors Dennard scalingbroke Multicore Era Single-core Era 3.5 GHz 3.4 GHz 740 KHz 2013 1971 2003 2004

32. Are multicores a long-term solution or just a stopgap?

33. Agenda • Who Hadi is • Course organization • Why alternative computing technologies • How we became and industry of new possibilities • Why we might become an industry of replacement • Possible alternative computing technologies • Quiz # 1

34. Modeling future multicoresQuantify the severity of the problem Predict the performance of best-case multicores • From 45 nm to 8 nm • Parallel benchmarks • Fixed power and area budget Transistor Scaling Model Single-Core Scaling Model Multicore Scaling Model Esmaeilzadeh, Belem, St. Amant, Sankaralingam, Burger, “Dark Silicon and the End of Multicore Scaling,”ISCA 2011

35. Transistor scaling modelFrom 45 nm to 8 nm [VLSI-DAT, 2010] [Dennard, 1974] [ITRS, 2010] Optimistic Scaling Model Historical Scaling Conservative Scaling Model Area 32× ↓ 32× ↓ 32× ↓ Power 8.3× ↓ 32× ↓ 4.5× ↓ Speed 5.7× ↑ 3.9× ↑ 1.3× ↑

36. Single-core model (45 nm) Power-Performance and Area-PerformancePareto Optimal Frontiers

37. Single-core scaling model From 45 nm to 8 nm Transistor Speed Scaling Factor Transistor Power Scaling Factor Single-core Scaling Model: Single-core Model × Transistor Scaling Model

38. Multicore scaling model From 45 nm to 8 nm Single Core Search Space (Scaled Area and Power Pareto Frontiers) Multicore Organization: CPU-Like, GPU-Like (# of HW Threads, Cache Sizes) Multicore Topology (Symmetric, Asymmetric, Dynamic, Composable) Microarchitectural Features (Cache and Memory Latencies, CPI, Memory Bandwidth) Constraints (Area and Power Budget) Application Characteristics (% Parallel, % Memory Accesses) Exhaustive search of multicore design space (Examine 800 design points for every technology node)

39. Multicore model (Amdahl’s Law)

40. Dark silicon 40%

41. Evaluation Setup • Applications: • 12 PARSEC Parallel Benchmarks • Baseline: • The best multicore design available at 45 nm • Constraints: • Driven from the best multicore design at 45 nm • Fixed Power Budget: 125 W • Fixed Area Budget: 111 mm2

42. 2013 18× 7.9× 3.7× 45 nm 10 years 8 nm 32 nm 22 nm 16 nm 11 nm Dark Silicon 1% 17% 36% 40% 51%

43. Industry of replacement? • Multicores are likely to be a stopgap • Not likely to continue the historical trends • Do not overcome the transistor scaling trends • The performance gap is significantly large • Radical departures from conventional approaches are necessary • Extract more performance and efficiency from silicon while preserving programmability • Explore other sources of computing

44. Agenda • Who Hadi is • Course organization • Why alternative computing technologies • How we became and industry of new possibilities • Why we might become and industry of replacement • Possible alternative computing technologies • Quiz # 1

45. Alternative computing technologies Human-basedComputing Biological Computing Approximate Computing Analog Computing Neuromorphic Computing Perpetual Computing Stochastic Computing

46. Approximate computingEmbracing error • Relax the abstraction of near-perfect accuracy in general-purpose computing • Allow errors to happen in the computation • Run faster • Run more efficiently

48. New landscape of computingPersonalized and targeted computing

49. Classes ofapproximate applications • Programs with analog inputs • Sensors, scene reconstruction • Programs with analog outputs • Multimedia • Programs with multiple possible answers • Web search, machine learning • Convergent programs • Gradient descent, big data analytics

50. Adding a third dimensionEmbracing Error Error