PCCAT Keynote 1jun12: We are already designing integrated circuits with more than a Billion transistors; and despite the faltering steps of Moore's law, we will be designing more than 10 times that within the next 5yrs. How will industry ever deliver a complete design in a reasonable time-frame today? The answer is Reuse and Hierarchy, both vitally important tools, yet neither academic topics in their own right.

1. Title: Making Molehills of Mountains
 Abstract:
 We are already designing integrated circuits with more than a Billion transistors; and despite the faltering
steps of Moore's law, we will be designing more than 10 times that within the next 5yrs. Yet even using High
Level Description Languages (HDLs) with Synthesis, designer productivity still only delivers the low thousands
of gates per day. How does industry ever deliver a complete design in a reasonable time-frame today? The
answer is Reuse and Hierarchy. You don't design everything from scratch, but strive to use as much as
possible of 'the last' design in the next. If you can hit 99%, then the 1B transistor opportunity, becomes a
10M transistor challenge ... still a large number, but much more manageable. Reuse is an indispensible part
of product design today, yet seldom gets the academic attention it deserves. This talk will examine reuse
today and its role in the pragmatic delivery of Electronic System products in the near and not-so-near future.
 1hr talk (Panel discussion later)
 Context
 The Postgraduate Conference for Computing: Applications and Theory (PCCAT) is intended to give postgraduate students from
the South West of the UK a flavour of presenting their research in a conference environment.
 The conference is held annually at a university in the South West, and is organised by postgraduate students. It has so far been
held twice, at the University of Exeter in 2010 and 2011, and will be held in 2012 at the University of Plymouth.
 http://www.pccat.ex.ac.uk/
1

2. 1v1
Prof. Ian Phillips
Principal Staff Eng’r,
ARM Ltd
ian.phillips@arm.com
Visiting Prof. at ...
PCCAT Conference
Uo.Plymouth
Contribution to Industry
6jun12
Award 2008
2

3. Our 21c World ...
 Statistics ...
 Population ~7,000,000,000
 Growth rate ~2%pa
 Life expectancy 60-80yr
... Mission: Celebrity, Leisure
3

4. Engineering in the UK ...
... Engineering made the world we live in; yet most people can’t see it !
4

5. The Pre-Engineered World (2,500 BC - 800 AD.)
 World Stats ...
 Population ~100K ->1M (Outnumbered!)
 Growth rate ~0.1%pa
 Life expectancy 30-40yr
... Mission: Survive and Grow
 Technology ...
 Low stone wall for a base,
 Wooden poles and rafters.
 Thatch, turf, or hides for roof.
 Timber split using 'wedges
 Sharp stones for cutting
 ... 3,500yrs of: “If it was good enough for my father’s, father’s,
father’s, father; its good enough for me!”
... Engineering brought mankind out
of the mud-hut !
5

6. Chronology of Science / Engineering Universe – 13.6Byr
Earth – 4.5Byr
 Cro-Magnon Man (Us!) – 35,000 yr ago
 ‘Developed’ from Homo-Sapien (Wise Human) 100,000 yr ago
 Mission: Survive Nature (1,000 generations)
 The Philosophers – 2,500-1,000 yr ago
 Pythagoras, Socrates, Plato, Aristotle, Archimedes, ...
 Mission: Understanding Nature
 The Scientists – 1,000-500 yrs ago
 Galileo, Descartes, (1000 ad)
 Electricity - William Gilbert (1600ad)
 Mission: Manipulation of Nature
 The Engineers – 260 yrs ago
 Industrial Revolution (1750: 8 gen’n)
 Year 0: Science Meets Exploitation
 Mission: Exploitation of Nature
... Economic (and Population) Explosion
Thomas Telford’s Iron Bridge (1778), Ironbridge, UK
6

7. The Industrial Revolution (1750)
 Exploitation of Nature
 Unleashing the Power of Science, by delivering it in ways that satisfied a
Volume Need ... We now call this Business.
 It began in the United Kingdom, then spread throughout Europe, North
America, and eventually the world.
 Major changes in agriculture, manufacturing, mining, transportation, and
technology
 Mechanisation of the textile industries,
 Development of iron-making techniques
 Trade expansion through canals, improved roads and railways.[5]
 Steam power, water wheels and powered machinery
 Profound effect on socio-economic and cultural conditions
... For the first time in history (13.6Byr), the living standards of the
masses of ordinary people underwent sustained growth
7

8. Our 21c World is ...
 Engineered Science ...
 Using a universal
Monetary System
 And significant Reuse of
Knowledge and Know-How
8

9. Science enables us to do ...
It is pretty clever to convert
a Stone into a Phone ...
 But we know it’s not Magic!
 Its just the measure of what
humans can achieve by
reusing the ingenuity of our
predecessors!
“If I have seen further it is
by standing on ye sholders
of Giants”
Isaac Newton 1676
“Reuse” Is the Elephant in the Room; The Cinderella at the Ball!
9

10. The Threshold of Magic 1: Clarke: Any sufficiently
advanced technology is
indistinguishable from magic.
 Everybody has a threshold, beyond which observed Functionality is
Indistinguishable From Magic1!
 Chemical Systems
 Biological Systems
 Economic Systems
 Electronic Systems
 The Incandescent Light:
is the
for most non-scientific,
but well educated people!
... Not Understand Technology is not a crime!
... The crime is not realising that most people don’t,
when you are the one to suffer as a consequence!
10

11. Antikythera c87BC ... Planet Motion Computer
Mechanical
Technology
• Inventor: Hipparchos (c.190 BC – c.120 BC).
Ancient Greek Astronomer, Philosopher and Mathematician.
• Single-Task, Continuous Time, Analogue Mechanical Computing (With backlash!)
See: http://www.youtube.com/watch?v=L1CuR29OajI
11

12. Computer: A Machine for Computing ...
Computing ...
... A general term for algebraic manipulation of data ...
Numerated Processed Data/
Phenomena Information
y=F(x,t,s) OUT (y)
IN (x)
... State and Time are normally factors in this.
 It can include phenomena ranging from human thinking to calculations
with a narrower meaning. Wikipedia
 Usually used it to exercise analogies (models) of real-world situations;
Frequently in real-time.
... No mention of Implementation Technology in this!
12

13. Orrery c1700 ... Planet Motion Computer
Mechanical
Technology
• Inventor: George Graham (1674-1751). English Clock-Maker.
• Single-Task, Continuous Time, Analogue Mechanical Computing (With backlash!)
13

14. Babbage's Difference Engine 1837
Mechanical
(Re)construction Technology
c2000
 The difference engine consists of a number of columns, numbered from 1 to N. Each column is able to store one decimal number. The only operation the engine
can do is add the value of a column n + 1 to column n to produce the new value of n. Column N can only store a constant, column 1 displays (and possibly prints)
the value of the calculation on the current iteration.
Computer for Calculating Tables: A Basic ALU Engine
14

15. “Enigma” c1940
Mechanical
Technology
Data Encryption/Decryption Computer
15

16. “Colossus” 1944
Valve/Mechanical
Technology
Code-Breaking Computer: A Data Processor
16

17. “Baby” 1947 (Reconstruction)
Valve/Software
Technology
General Purpose, Quantised Time and Data, (Digital) Electronic Computing
17

18. Analogue Computer – AKAT c1960
Transistor
Technology
General Purpose, Continuous Time, Approximate (Analogue) Electronic Computing
18

19. The Pinnacle of Computing Technology?
Integrated Circuit
& Software
Technology
19

20. ... Or the Smart Gadget ...
Electronic System
Technology
20

21. Or the Cloud ...?
Electronic System
Technology
21

22. Evolution of Radio
Tele-Verta Radio
4 Valves
1 Rectifier Valve
BTH c1945
Crystal Set
1 Diode Evoke DAB Radio
c1925 100 M Transistors
2-3 Embedded Processors
Bush Radio c2005
7 Transistors
1 Diode
c1960
Ian’s ‘Span’
22

23. Radio as Computation ...
Integrated Circuit
Transistor
Valve
Technology
Vi
Vrf=Vi*100
Vrf Vro='Bandpass'(Vif*1000)
Vif
Vro
Vif=Vrf*Vlo
Vlo
Vlo=Cos(t*1^6)
Single-Task, Continuous Time, Approximate (Analogue) Electronic Computing
23

24. Computing is
24

25. Its Products that Make Money
 21c Businesses have to be
 Selling things that People (End-Customers) want to buy.
 Globalisation makes them Focus on Core Competencies
 Objective: Outsource ‘everything else’
 Customers, Competition, Operations and Investors are Global
 Nationality: has little meaning
 Business needs
 End-Customers buy Functionality not Technology
 Technologies enable Product Options ..but..
 Business-Models make Money
 New Products are
 Design is a Cost/Risk to be Minimised
 Technology (HW, SW, Mechanics, Optics, etc) is (just) a
means to a Product end!
 New Technology increases Cost/Risk ... But not always Value
... Reuse Minimises the Risk and Cost of deploying any Technology
25

26. Manipulating Atomic Properties ...
 Electronic Technology is ..
...The Most Exciting thing mankind has created in our 35kyr history!
Early Electronics The First Transistor (1947) Modern Transistor
~70 yrs
... And it has all happened within the span of one life-time!
26

27. Moore’s Law: c1965
 “Moore's Law” was coined by Carver Mead in 1970, from Gordon
Moore's article in Electronics Magazine 19 April 1965 "Cramming
more components onto integrated circuits“.
“The complexity for minimum
component costs has increased at a rate
of roughly a factor of two per year ...
Certainly over the short term this rate can be
expected to continue, if not to increase. Over
the longer term, the rate of increase is a bit
more uncertain, although there is no reason to
believe it will not remain nearly constant for at
least 10 years. That means by 1975, the number
of components per integrated circuit for
minimum cost will be 65,000. I believe that such
a large circuit can be built on a single wafer”
Gordon Moore, Founder of Intel
In 1965 he was designing ICs with ~50 transistors!
Moore’s Law has held for ~50 years ... Taking us to 100B transistor ICs
27

28. Approximate Process Geometry
28
1um
10um
10nm
100um
100nm
Transistors/Chip (M)
Moore’s Law ...
X
ITRS’99
Transistor/PM (K)

29. All Exponentials Must End ...
130nm
 Growing opinion that 14 or 7nm will be
the smallest yieldable node ... Ever!
90nm  Just 3-4 gen. (5-8yr) to the
end of Planar Scaling
30nm  Only things on
the drawing
board today ...
14nm
... can get into the
last of the of planar chips! 7nm
 Its the end-of-the-road for
‘promising technologies’ !
 Clean-Sheet Synthesis
 Scalable Processor Arrays
 Formal Design ...The future lies
 Top-Down Design with Hybrid, Architectures
29

30. All Exponentials Must End ...
130nm
 Growing opinion that 14 or 7nm will be
the smallest yieldable node ... Ever!
90nm  Just 3-4 gen. (5-8yr) to the
end of Planar Scaling
30nm  Only things on
the drawing
board today ...
14nm
... can get into the
last of the of planar chips! 7nm
 Its the end-of-the-road for
‘promising technologies’ !
 Clean-Sheet Synthesis
 Scalable Processor Arrays
 Formal Design ...The future lies
 Top-Down Design with Hybrid, Architectures
30

31. … 3D will Maintain the Prediction!
 Die Stacking today Hybrid Technology
10x Sampling
 Die-Integration PV - 500nm Ge
RF - 300nm GaAs
tomorrow 8x Sampling CPU- 90nm Si CMOS
DRAM - 20nm Si FIN-MOS
Active Carrier 300nm Si CMOS
10 stack 1.6 mm
4x Transfer
to Production
3x Production
8 stack 1.6 mm
4 stack 1.4 mm ... This is a disconnect for
Moore’s Silicon Scaling Law, but not
3 stack 1.2 mm
for his ‘density’ predictions!
31

32. Moore's Real Law: x2 Functionality Every 18mth!
 Cascade of Technologies supporting Functional growth ...
1012
Functional Density (units)
1010
106
102
Electronic era: System era:
1975-2005 2003-2030
100
1960 1980 2000 2020
... It Started with Wood ⇒ Stone ⇒ Bronze ⇒ Iron
32

33. Moore’s Productivity Law ...
10nm
Approximate Process Geometry
100nm
1um 2x in 18mth
Transistors/Chip (M)
Transistor/PM (K)
16% in 18mth
10um
100um
ITRS’99
33

34. Where did the Productivity Gap go?
Reuse Happened !
 <1990 chip design was entire ...
 Moore’s Law was handled by ever Bigger Teams and ever Faster Tools
 With Improved Productivity through HDL and Synthesis
... I was a chip designer in 1975; and did it all, myself, in 3mth (1k gates!)
 >1995 reuse quietly entered the picture ...
 Circuit Blocks
 CPUs (and Software) ... With
 External IP Supporting
Methodology!
 Up-Integration
(Incl. Software)
 Chip Reuse (ASSP)
... Delivering Productivity, Quality and Reliability
... Birth of HW/SW IP Companies (eg ARM)
... But brought about Commoditisation of Silicon (and FABs) !
34

35. How Much Reuse Do We Do?
 Mobile Products have 500m gate SoCs / 500m lines of code
 Doubling every 18mth
 Designer Productivity: is just 100-1000 Gates(Lines)/day
 That is tested, verified, incorporated gates/lines
 That’s 2,500-25,000 p.yrs to design! (Un-Resourceable)
 Typically ‘Designs’ have just 50-100 p.yr available ...
 That’s just ~0.5% New (>99% Reuse!)
 Not Viable to do a clean-sheet product design (Nor has been since ~1995)
 The core HW/SW is only a part of a Product
 There’s all of the other Components and Sub-Systems
 There’s the IO systems (RF, Audio, Optical, Geo-spatial, Temporal)
 There’s the Mechanical
 There’s the Reproduction (Factory)
 There's the Business Model (Cash-flow, Distribution, Legal)
 There’s the Support (Repair, Installation, Maintenance, Replacement)
35

36. How Much Reuse Do We Need?
 Design Tools (across all Product Disciplines) underpin this ...
 Reuse of Modules and Components
 Reuse of Existing Code and Circuits
 Sharing Methodology
 Sharing Architecture
 Creating Tools to Accelerate Methodology and Repeatability
 Design For “x” (DFx) is Design For up-stream Deployment
 Includes DFR (Design For Reuse)
 A significant part is (and will remain) Knowledge based ...
 The Designer has done similar work before
 The Team has Collective experience
 The Company has experience and a customer base
 The nature of the Design Engineering Role is ...
 To create Order out of Chaos
 To apply state-of-the-art and knowledge; to create a Producible Product
36

37. The Technology in an iConic Product ...
37

38. ... Cool Design
38

39. ... Design happens at Many Levels ...
39

40. Inside the Case ...
Down 1-Level: Modules
iPhone 4's vibrator motor. rear-facing 5 MP camera with
720p video at 30 FPS, tap to
focus feature, and LED flash.
40 Source ... http://www.ifixit.com

41. Inside the Case ...
Down 1-Level:
Modules
The Control Board.
41 Source ... http://www.ifixit.com

42. Inside The Control Board (b-side)
Down 2-Levels: Sub-Assemblies
 Visible Design-Team Members ...
 Samsung (flash memory) - (ARM Partner)
 Cirrus Logic (audio codec) - (ARM Partner)
 AKM (Magnetic Sensor)
 Texas Instruments (Touch Screen Controller and mobile DDR) - (ARM Partner)
 Invisible Design-Team Members ...
 Software Tools, OS & Drivers, GSM Security; Graphics, Video and Sound ...
 Manufacturing, Assembly, Test, Certification ...
42 Source ... http://www.ifixit.com

43. Inside The Control Board (a-side)
Down 2-Levels: Sub-Assemblies
 Visible Design-Team Members...
 A4 Processor, specified by Apple, designed and manufactured by Samsung ...
 The central unit that provides the iPhone 4 with its GP computing power.
 Reported to contain ARM A8 600 MHz CPU (other ARM CPUs and IP)
 ST-Micro (3 axis gyroscope) - (ARM Partner)
 Broadcom (Wi-Fi, Bluetooth, and GPS) - (ARM Partner)
 Skyworks (GSM)
 Triquint (GSM PA)
 Infineon (GSM Transceiver) - (ARM Partner)
GPS
Bluetooth,
EDR &FM
43 Source ... http://www.ifixit.com

44. The A4 SIP Package (Cross-section)
Memory
‘Package’
2 Memory Dies
Glue Processor SOC Die
4-Layer Platform
Package’
Down 3-Levels: IC Packaging
 The processor is the centre rectangle. The silver circles beneath it are solder balls.
 Two rectangles above are RAM die, offset to make room for the wirebonds.
 Putting the RAM close to the processor reduces latency, making RAM faster and cuts power.
 Unknown Mfr (Memory)
 Samsung/ARM (Processor)
 Unknown (SIP Technology)
44 Source ... http://www.ifixit.com

45. Lots and Lots of Designers ...
 159 Tier-1 Suppliers ...
 Thousands of Design Engineers
 10’s of thousands of Engineers
 Globally
... Hundreds more Tier-2
suppliers (Including ARM)
45

46. Putting ‘Smart’ into Electronic Systems …
46

47. The ARM RISC-Processor Core
ADDR[31:0]
Address
Incrementer Scan
Debug
Address Register Incrementer Control
P CFGBIGEND
C CLK
CLKEN
PC Update WRITE
Register Bank Instruction SIZE[1:0]
Decoder
Decode
Stage nIRQ
nFIQ
A A B Instruction nRESET
Multiplier B Decompression and ABORT
L B
U u u TRANS
B s s
PROT
u Barrel Control LOCK
s Shifter Logic
CPnOPC
CPnCPI
Write Data Read Data CPA
32 Bit ALU CPB
Register Register
WDATA[31:0] RDATA[31:0]
47

48. The ‘Lego-Brick’ Chip-Design Concept
nVidea Tegra3
ARM
ARM
ARM ARM
ARM ARM
48

49. More and More Systems on a Chip
Users require a pocket ‘Super-Computer’ ...
 Silicon Technology Provides a few-Billion raw transistors ...
 ARM’s IP makes it Practical to utilise them ...
• 10 Programmable Processors
• 4 x A9 Processors (2x2):
• 4 x MALI 400 Fragment Proc:
• 1 x MALI 400 Vertex Proc.
• 1 x MALI Video CoDec
• Software Stacks, OS’s and
Design Tools/
• ARM Technology gives
chip/system designers a
good start. Design Reuse ...
• Improves Productivity
• Improves TTM
• Improves Quality/Certainty
49

50. ARM Technology
 Electronic System products incorporate
more and more ARM technology –Processor,
Multimedia
and Software IP
Processor IP – Design of the
brain of the chip
Physical IP – Design of the building
blocks of the chip
Software & Development tools
... 800 Partners; 600 Licences in 200 Companies
... Millions of developers; Billions of users
50

51. The World’s Favourite IP Provider
 1990 - "A barn in Cambridge"
 12 engineers, in Cambridge
 No Revenue, No Patents
 Cash from Apple & VLSI
 Spin-out of Acorn UK ...
 BBC Computers in Schools (1981)
 Roots in Uo.Cambridge (c1975)
... A Dream to become the Global
Standard for Embedded CPUs
 2012 - "The worlds leading IP Product"
 Powering >90% of the Smart Electronic Systems in the world
 7B CPUs shipped in 2011 ... Growth ~25%pa; 40B total (>50x all PCs!)
 FTSE 100 company: Revenue ~£491M, PBT ~37%, R&D ~30%
 Cambridge HQ: 25 offices/labs 2000 people ww (850 in the UK)
 95% revenue is foreign earnings
51

52. An Irresistible Societal Trend ...
 Electronic Systems ...
+ Get Smarter + Get Smaller/Cheaper
The Internet of Things
+ Get Pervasive + Talk to One Another
100 Billion
+ Need no Attention + Work Better
... Cease To Be Noticed !
Mobile Internet
10 Billion
Desktop
Units
Internet
PC 1 Billion
100M
Mini
2nd Era
Mainframe 10M
1M 1st Era Cost
1960 1970 1980 1990 2000 2010 2020
52

53. Expectations of Tomorrow’s Consumer
Natural, Intuitive User
Interfaces Continuous Connectivity
Personalized
Experience
Ultra High Resolutions
Displays Standout Battery Life
“Always On, Always Connected”
53

54. High Resolution Natural Displays
Ultra Thin and
Flexible Displays
High resolution,
interactive wall
size displays
Augmented reality
and visual
computing
54

55. Natural Speech Recognition
Siri – Best in class today…
…but multi language and accents are still challenging
55

56. Heterogeneous Computing
 Multi-core solutions
 Close coupling of different
types of processors cores
 Shared memory architecture
 Match task to most efficient
core
User Interface Augmented Reality Image Recognition/Processing Advanced Gaming
56

57. big.LITTLE Processing
 Uses the right processor for the right job
 Redefining the efficiency/performance trade-off
 70% energy savings on common workloads
 Flexible and transparent to apps
big LITTLE
“Demanding tasks” Interrupt Control “Always on, always
Cortex-A15
connected tasks”
MPCore Cortex-A7
CPU CPU MPCore
CPU CPU
L2 Cache L2 Cache
CCI-400 Coherent Interconnect

58. 2015 Mobile Device Trends
Intuitive and Natural User Interfaces
More pixels, greater fidelity
Continued Innovation on Energy Efficiency
Protection of your Data
Always on, Always Connected
Leading Innovation in Computing Devices
Delivered through a single Architecture and a broad partnership
58

59. Innovation & Efficiency Underpins It All
59

60. ARM – Architecture for the Digital World
150+
billion
chips cumulative
in 2020
30+
billion
chips to date
1998 2012 2020
60

61. Viewing The Mountain ...
 A single 40nm SoC can have a Billion or more Gates on it
 Embedded Products will embody a Billion Lines of Code
 All Multi-level State Machines are too complex to simulate
 Imperfection happens: Impacting Reliability and Robustness
 Nobody writes HDL/Code without errors
 All Systems are too complex to Test
 What happens after 20nm?
 Design teams are finite  Notate
 Market windows close  Quantify
 Quality is Qualitative  Automate
 Power Dissipation!  Consolidate
 2nd is for Losers  Learn Generic Lessons
 Productivity!  Think-out of Consequences
 ... ... Build Shoulders for others to Stand On
61

62. The 21C will be what YOU Make It ...
Thankyou
for
Listening
“Any sufficiently advanced technology is indistinguishable from magic! Arthur C. Clarke.
62

63. Reading & References
 Electronics 2015: Making a Visible Difference (Referred)
 DTI EIGT Report, HMG URN 04/1812, 2004.
 Engineering UK 2009 (and 2011): The state of engineering (Referred)
 EngineeringUK (ex Engineering Council), 2009 and 2011.
 The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail (Disruptive Tech.)
 by Clayton M. Christensen: HBS Press, 1997
 Open Innovation: The New Imperative for Creating and Profiting from Technology (Research in 21C)
 by Henry William Chesbrough : HBS Press, 2003
 The World Is Flat (Globalisation)
 by Thomas L. Friedman: Penguin, 2005
 Staying Power (Business)
 by Michael Cusumano: Oxford, 2010
 A Short History of Nearly Everything (A different view on what we know)
 by Bill Bryson: Black Swan, 2003
 The Voyages of the Beagle (Scientific Observation) – Free on-line
 By Charles Darwin,1860
 An Essay on the Principles of Population (Natural Competition) – Free on-line
 By Thomas Malthus,1789
63