vlsi design automation

30
Calcolatori Elettronici – Ing. Informatica 1 VLSI Design Automation VLSI Design Automation

Upload: others

Post on 24-Jan-2022

19 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 1

VLSI Design AutomationVLSI Design Automation

Page 2: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 2

OutlineOutlineTechnology trendsVLSI Design flow (an overview)

Page 3: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 3

IC ProductsIC Products Processors

CPU, DSP, Controllers Memory chips

RAM, ROM, EEPROM Analog

Mobile communication,audio/video processing

ProgrammablePLA, FPGA

Embedded systemsUsed in cars, factoriesNetwork cards

System-on-chip (SoC)

Page 4: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 4

Integrated Circuit RevolutionIntegrated Circuit Revolution

1972: Intel 4004 MicroprocessorClock speed: 108 KHz# Transistors: 2,300# I/O pins: 16Technology: 10μm

Page 5: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 5

Integrated Circuit RevolutionIntegrated Circuit Revolution

2000: Intel Pentium 4 ProcessorClock speed: 1.5 GHz# Transistors: 42 millionTechnology: 0.18μm CMOS

Page 6: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 6

Integrated Circuit RevolutionIntegrated Circuit Revolution

2006: Intel Core 2 DuoClock speed: 3.73 GHz# Transistors: 1 billionTechnology: 65nm CMOS

Page 7: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 7

Integrated Circuit RevolutionIntegrated Circuit Revolution

2005: Sun UltraSpartc T18 cores, 4 threads per coreClock speed: 1.2 GHz# Transistors: 300 millionTechnology: 90nm CMOS

Page 8: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 8

Moore’s LawMoore’s Law Gordon Moore predicted in 1965 that the number of transistors that

can be integrated on a die would double every 18 months.

Page 9: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 9

Semiconductor GrowthSemiconductor Growth

Page 10: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 10

Processor Power (Watts)Processor Power (Watts)

Page 11: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 11

Intel Microprocessor PerformanceIntel Microprocessor Performance

Page 12: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 12

Device ComplexityDevice Complexity Exponential increase in device

complexityIncreasing with Moore's law (or faster)!

Require exponential increases in design productivity

We have exponentially more transistors!We have exponentially more transistors!

Page 13: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 13

Heterogeneity on ChipHeterogeneity on ChipGreater diversity of on­

chip elementsProcessors Software Memory Analog

More transistors doing different things!More transistors doing different things!

Page 14: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 14

Stronger Market PressuresStronger Market PressuresTime–to-market

Decreasing design windowLess tolerance for design

revisions

Page 15: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 15

How Are We Doing?How Are We Doing?

Role of EDA: close the productivity gap

Log

ic t

ran

sist

ors

per

chip

(K

)

ProductivityTrans. / Staff . M

onth

10

100

1,000

10,000

100,000

1,000,000

10,000,000

100,000,000

Tr./S.M

10

100

1,000

10,000

100,000

1,000,000

10,000,000

Logic Tr./C

hip

1981

1985

1989

1993

1997

2001

2005

2009

58% / Yr. compoundcomplexity growth rate

21% / Yr. compoundproductivity growth rate

Productivitygap

Page 16: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 16

Evolution of Design MethodologyEvolution of Design MethodologyWe are now entering the era of block-based

design

ASIC/ASSPDesign

System-BoardIntegration

YesterdayBus Standards,

Predictable, Preverified

TodayVSI Compatible Standards,

Predictable, Preverified

IP/BlockAuthoring

System-ChipIntegration

Page 17: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 17

Evolution of SoC PlatformsEvolution of SoC PlatformsGeneral-purposeScalable RISCProcessor• 50 to 300+ MHz• 32-bit or 64-bit

Library of DeviceIP Blocks• Image coprocessors• DSPs• UART• 1394• USB

Scalable VLIWMedia Processor:• 100 to 300+ MHz• 32-bit or 64-bit

Nexperia™System Buses• 32-128 bit

2 Cores: Philips’ Nexperia PNX8850 SoC platform for High-end digital video (2001)

Page 18: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 18

What’s Happening in SoCs?What’s Happening in SoCs? Technology: no slow-down in sight!Faster and smaller transistors: 90 → 65 → 45 → 32 nm

… but slower wires, lower voltage, more noise!80% or more of the delay of critical paths will be due to

interconnects Design complexity: from 2 to 10 to 100 cores!

Design reuse is essential…but differentiation/innovation is key for winning on the

market! Performance and power:

Performance requirements keep going up…but power budgets don’t!

Page 19: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 19

Communication ArchitecturesCommunication ArchitecturesShared bus

Low areaPoor scalabilityHigh energy consumption

Network-on-ChipScalability and modularityLow energy consumptionIncrease of design complexity

Shared bus

IP IP IP

IP IP IP

IP IP IP IP

IP IP IP IP

IP IP IP IP

IP IP IP IP

Page 20: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 20

Intel’s TeraflopsIntel’s Teraflops 100 Million transistors 80 cores, 160 FP engines Teraflops perf. @ 62 Watts On-die mesh network Power aware design

Page 21: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 21

IC Design StepsIC Design StepsSpecificationsSpecifications High-level

DescriptionHigh-level

DescriptionFunctionalDescriptionFunctionalDescription

BehavioralVHDL, C

StructuralVHDL

Page 22: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 22

IC Design StepsIC Design Steps

Packaging Fabri-cation

PhysicalDesign

TechnologyMapping

Synthesis

SpecificationsSpecifications High-levelDescriptionHigh-level

DescriptionFunctionalDescriptionFunctionalDescription

Placed& RoutedDesign

Placed& RoutedDesign

X=(AB*CD)+ (A+D)+(A(B+C))Y = (A(B+C)+AC+ D+A(BC+D))

Gate-levelDesign

Gate-levelDesign

LogicDescription

LogicDescription

Page 23: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 23

Circuit ModelsCircuit ModelsA model of a circuit is an abstraction

A representation that shows relevant features without associated details

Circuit Model(few details)

Circuit Model(few details)

Circuit Model(many details)Circuit Model(many details)SynthesisSynthesis

Page 24: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 24

Model ClassificationModel Classification

Page 25: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 25

Levels of AbstractionLevels of AbstractionArchitectural

A circuit performs a set of operation, such as data computation or transferHDL models, Flow diagrams, …

LogicA circuit evaluate a set of logic functions

FSMs, Schematics, …

GeometricalA circuit is a set of geometrical entities

Floor plans, layouts, ...

Page 26: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 26

Levels of AbstractionLevels of Abstraction

Design consists of refining the abstract specification of the architectural model into the detailed geometrical-level model

…PC = PC + 1;Fetch(PC);Decode(Inst);...

Page 27: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 27

Views of a ModelViews of a ModelBehavioral

Describe the function of a circuit regardless of its implementation

StructuralDescribe a model as an interconnection of

componentsPhysical

Relate to the physical object (e.g., transistors) of a design

Page 28: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 28

The Y-chartThe Y-chart

Architectural-level

Logic-level

Geometrical-level

Behavioral-view Structural-view

Physical-view Gajski and Kuhn’s Y-chart(Silicon Compilers, Addison-Wesley, 1987)

Page 29: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 29

The Y-chartThe Y-chartBehavioral-view Structural-view

Physical-view

Architecturallevel

Logic level

Geometricallevel

…PC = PC + 1;Fetch(PC);Decode(Inst);...

MULT

ADDRAM

CTRL

S0

S2

S3 S1

Page 30: VLSI Design Automation

Calcolatori Elettronici – Ing. Informatica 30

SynthesisSynthesis

Architectural-level

Logic-level

Geometrical-level

Behavioral-view Structural-view

Physical-view

High-level synthesis(or architectural synthesis)

Logic synthesis

Physical design

Assignment to resources Interconnection Scheduling

Interconnection of istances of library cells (technology mapping)

Physical layout of the chip (placement, routing)