Download presentation
Presentation is loading. Please wait.
Published byDianna Griffith Modified over 10 years ago
1
® 1 Exponential Challenges, Exponential Rewards— The Future of Moore’s Law Based on lecture of Shekhar Borkar Intel Fellow Circuit Research, Intel Labs
2
2 ISSCC 2003— Gordon Moore said… “No exponential is forever… But We can delay Forever”
3
3 Goal: 1TIPS by 2010 Pentium® Pro Architecture Pentium® 4 Architecture Pentium® Architecture 486 386 286 8086 How do you get there?
4
4 Transistors Scaling Will high K happen? Would you count on it?
5
5 Technology Scaling GATE SOURCE BODY DRAIN Xj Tox D GATE SOURCE DRAIN Leff BODY Dimensions scale down by 30% Doubles transistor density Oxide thickness scales down Faster transistor, higher performance Vdd & Vt scaling Lower active power Technology has scaled well, will it in the future?
6
6 Gate Oxide is Near Limit 70 nm Si 3 N 4 CoSi 2 130nm Transistor Will high K happen? Would you count on it? GATE SOURCE BODY DRAIN Tox GATE SOURCE DRAIN 70 nm BODY
7
7 3D-Gate Transistor
8
8 Transistor Integration Capacity On track for 1billion transistor integration capacity
9
9 35 Years of Microprocessor Trend C Moore, Data Processing in ExaScale-Class Computer Systems, Salishan, April 2011
10
10 Transistor Integration Capacity
11
11 Transistor Integration Capacity
12
12 Transistor Integration Capacity
13
13 Transistor Integration Capacity
14
14 Exponential Challenge #1
15
15 Is Transistor a Good Switch? On I = ∞ I = 0 Off I = 0 I ≠ 0 I = 1ma/u I ≠ 0 Sub-threshold Leakage
16
16 Sub-threshold Leakage Sub-threshold leakage increases exponentially Assume: 0.25 m, I off = 1na/ 5X increase each generation at 30ºC
17
17 Leakage Power Leakage power limits Vt scaling A. Grove, IEDM 2002
18
18 The Power Crisis
19
19 How Power Should Have Scaled A. Danowitz et al. CPU DB: Recording Microprocessor History. ACMQueue Processors, vol. 10, issue 4, pp1-18. 2012
20
20 Exponential Challenge #4
21
21 Impact on Path Delays Path Delay Path delay variability due to technological variations Impacts individual circuit performance and power Optimize each circuit for performance and power Delay Probability Due to variations in: Vdd, Vt, and Temp
22
22 Impact on Path Delays Path Delay Path delay variability due to technological variations Impacts individual circuit performance and power Optimize each circuit for performance and power Delay Probability Due to variations in: Vdd, Vt, and Temp How many silicon atoms (111pm) have on transistor channel (20nm)? 3D transistor is a solution?
23
23 Shift in Design Paradigm From deterministic design to probabilistic and statistical design From deterministic design to probabilistic and statistical design –A path delay estimate is probabilistic (not deterministic) Multi-variable design optimization for Multi-variable design optimization for – Parameter variations – Active and leakage power – Performance
24
24 Exponential Challenge #6
25
25 Exponential Costs G. Moore ISSCC 03 Litho Cost www.icknowledge.com FAB Cost $ per Transistor $ per MIPS
26
26 Some Implications Tox scaling will slow down—may stop? Tox scaling will slow down—may stop? Vdd scaling will slow down—may stop? Vdd scaling will slow down—may stop? Vt scaling will slow down—may stop? Vt scaling will slow down—may stop? Approaching constant Vdd scaling Approaching constant Vdd scaling Energy/logic op will not scale Energy/logic op will not scale
27
27 The Terascale Dilemma Many billion transistor integration capacity will be available Many billion transistor integration capacity will be available – But could be unusable due to power Logic transistor growth will slow down Logic transistor growth will slow down Transistor performance will be limited Transistor performance will be limitedSolutions Low power design techniques Low power design techniques Improve design efficiency Improve design efficiency
28
28 Exponential Challenge #5
29
29 Platform Requirements 0 500 1000 1500 2000 2500 3000 PC towerMini tower tower Slim lineSmall pc System Volume ( cubic inch) Shrinking volume Quieter Yet, High Performance 0 0.5 1.0 1.5 050100150200 Power (W) Thermal Budget ( o C/W) 0 25 50 75 Heat-Sink Volume (in 3 ) Projected Heat Dissipation Volume Projected Air Flow Rate Pentium ® III 100 250 Thermal Budget Air Flow Rate (CFM) Pentium ® 4 Thermal budget decreasing Higher heat sink volume Higher air flow rate
30
30 Active Power Reduction SlowFastSlow Low Supply Voltage High Supply Voltage Logic Block Freq = 1 Vdd = 1 Throughput = 1 Power = 1 Area = 1 Pwr Den = 1 Vdd Logic Block Freq = 0.5 Vdd = 0.5 Throughput = 1 Power = 0.25 Area = 2 Pwr Den = 0.125 Vdd/2 Logic Block Multiple Vdd Throughput oriented design
31
31 Design & Arch Efficiency Employ efficient design & Architectures
32
Improve Arch Efficiency ST Wait for Mem MT1 Wait for Mem MT2 Wait MT3 Single Thread Multi-Threading Thermals & Power Delivery designed for full HW utilization Multi-threading improves performance without impacting thermals & power delivery Computer Architecture: A Quantitative Approach (Hennessy;Patterson, 2011)
33
33 Increase on-die Memory Large on die memory provides: 1.Increased Data Bandwidth & Reduced Latency 2.Hence, higher performance for much lower power
34
34 Chip Multi-Processing Keynote presentation (L. Benini, RSP 2010).
35
35 Chip Multi-Processing C1C2 C3C4 Cache Multi-core, each core Multi-threaded Shared cache and front side bus Each core has different Vdd & Freq Spreading hot spots Lower junction temperature
36
36 Example (Itanium Tukwila)
37
37 Example (Itanium Tukwila) 30 MBytes cache 130 Watts
38
38 Example (Itanium Tukwila)
39
39 What the Cores Will look like?
40
40 What the Cores Will look like?
41
41 What the Cores Will look like?
42
42 What the Cores Will look like? clocks run with the same frequency but unknown phases
43
43 What the Cores Will look like?
44
44 What the Cores Will look like? Intelligent redistribution workload Improvement of energy efficiency Multiple functionalities
45
45 What the Cores Will look like? Several interconnection possibilities Mesh Ring
46
46 Tera-Scale RMS - Recognition, Mining and Synthesis
47
47 Tera-Scale
48
48 Tera-Scale
49
49 Tera-Scale
50
50 The Exponential Reward Speculative, OOO Era of Instruction LevelParallelism Super Scalar 486 386 286 8086 Era of Pipelined Architecture Multi Threaded Era of Thread & ProcessorLevelParallelism Special Purpose HW Multi-Threaded, Multi-Core
51
51 Summary—Delaying Forever Terascale transistor integration capacity will be available - Power and Energy are the barriers Terascale transistor integration capacity will be available - Power and Energy are the barriers Variations will be even more prominent - shift from Deterministic to Probabilistic design Variations will be even more prominent - shift from Deterministic to Probabilistic design Improve design efficiency Improve design efficiency Exploit integration capacity to deliver performance in power/cost envelope Exploit integration capacity to deliver performance in power/cost envelope
52
52 1. Discuta um problema associados a integração dos dispositivos 2. Comente a afirmação: - “A redução do tamanho dos transistores muda o paradigma de avaliação de consumo de energia e tempo de execução de determinístico para probabilístico” 3. Porque o consumo de energia estático é tão problemático para as tecnologias futuras? 4. Porque a redução da voltagem é um dos principais elementos a tratar para reduzir o consumo de energia? 5. Como um sistema com várias alimentações pode contribuir para a redução do consumo de energia? Qual o efeito sobre o tempo de execução? Exercícios
53
53 6. Faça uma ilustração que mostre como um programa multi- thread pode ocupar melhor os recursos de um sistema, reduzindo o gargalo de comunicação com a memória 7. Qual o motivo do percentual de memória interno a um circuito integrado passar de 50% nos processadores atuais? 8. Dada a limitação do escalamento, o que pode ser feito para continuar o crescente aumento do desempenho das máquinas? 9. Quais as tendências em termos de computação (cores), infra-estrutura de comunicação e armazenamento para os próximos processadores? Exercícios
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.