00 introduction
TRANSCRIPT
![Page 1: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/1.jpg)
cse141: Introduction to Computer Architecture
Steven SwansonHung-Wei Tseng
1
![Page 2: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/2.jpg)
Today’s Agenda
• What is architecture?• Why is it important?• At the highest level, where is architecture today?
Where is it going?• What’s in this class?
2
![Page 3: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/3.jpg)
What is architecture?
• How do you build a machine that computes?
• Quickly, safely, cheaply, efficiently, in technology X, for application Y, etc.
Civilization advances by extending the number of important
operations which we can perform without thinking about them.
-- Alfred North Whitehead
![Page 4: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/4.jpg)
Orientation
The internet
![Page 5: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/5.jpg)
Orientation
The internet
![Page 6: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/6.jpg)
OrientationSystem Bus
(PCI)IO
Power
Memory
Power
Memory
Memory MemoryArchitecture begins about here.
![Page 7: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/7.jpg)
OrientationSystem Bus
(PCI)IO
Power
Memory
Power
Memory
Memory MemoryArchitecture begins about here.
![Page 8: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/8.jpg)
![Page 9: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/9.jpg)
![Page 10: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/10.jpg)
You are here
![Page 11: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/11.jpg)
You are here
cse141
![Page 12: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/12.jpg)
The processors go here…
![Page 13: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/13.jpg)
The processors go here…
![Page 14: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/14.jpg)
Abstractions of the Physical World…
Physics/Materials Devices Micro-architecture ArchitecturesProcessors
![Page 15: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/15.jpg)
Abstractions of the Physical World…
Physics/Materials Devices Micro-architecture ArchitecturesProcessors
This Coursecse241a/ECE dept
Physics/Chemistry/
Material science
![Page 16: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/16.jpg)
…for the Rest of the System
Architectures
JVM
Processor
AbstractionCompilers Languages
Software
Engineers/Applications
![Page 17: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/17.jpg)
…for the Rest of the System
Architectures
JVM
Processor
AbstractionCompilers Languages
Software
Engineers/Applications
cse130cse121 cse131 cseEverythingElse
![Page 18: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/18.jpg)
Why study architecture?
11
• As CEs or CSs you should understand how computers work• Processors are the basis for everything in CS (except theory)• They are where the rubber meets the road.
• Performance is important• Faster machines make applications cheaper• Understanding hardware is essential to understanding how
systems behave
• It’s cool!• Microprocessors are among the most sophisticated devices
manufactured by people• How they work (and even that they work) as reliably and as
quickly as they do is amazing.
• Architecture is undergoing a revolution• The future is uncertain• Opportunities for innovation abound.
![Page 19: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/19.jpg)
Performance and You!
• Live Demo
12
![Page 20: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/20.jpg)
Processor are Cool!
• Chips are made of silicon• Aka “sand”• The most adundant element in the
earth’s crust.• Extremely pure (<1 part per billion)• This is the purest stuff people make
![Page 21: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/21.jpg)
Building Chips
![Page 22: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/22.jpg)
Building Chips
• Photolithography
Silicon Wafer
![Page 23: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/23.jpg)
Building Chips
• Photolithography
Silicon Wafer Silicon WaferSiO2
Grow silicon dioxide
![Page 24: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/24.jpg)
Building Chips
• Photolithography
Silicon Wafer Silicon WaferSiO2
Grow silicon dioxideSilicon Wafer
SiO2Resist
Apply photo resist
![Page 25: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/25.jpg)
Building Chips
• Photolithography
Silicon Wafer Silicon WaferSiO2
Grow silicon dioxideSilicon Wafer
SiO2Resist
Apply photo resistSilicon Wafer
SiO2Resist
Mask Mask
Expose to UV
![Page 26: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/26.jpg)
Building Chips
• Photolithography
Silicon Wafer Silicon WaferSiO2
Grow silicon dioxideSilicon Wafer
SiO2Resist
Apply photo resistSilicon Wafer
SiO2Resist
Mask Mask
Expose to UV
Silicon WaferSiO2
Patterned resist
![Page 27: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/27.jpg)
Building Chips
• Photolithography
Silicon Wafer Silicon WaferSiO2
Grow silicon dioxideSilicon Wafer
SiO2Resist
Apply photo resistSilicon Wafer
SiO2Resist
Mask Mask
Expose to UV
Silicon WaferSiO2
Patterned resistSilicon Wafer
Etch SiO2
![Page 28: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/28.jpg)
Building Chips
• Photolithography
Silicon Wafer Silicon WaferSiO2
Grow silicon dioxideSilicon Wafer
SiO2Resist
Apply photo resistSilicon Wafer
SiO2Resist
Mask Mask
Expose to UV
Silicon WaferSiO2
Patterned resistSilicon Wafer
Etch SiO2Silicon Wafer
Met
Deposit metal
![Page 29: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/29.jpg)
Building Chips
• Photolithography
Silicon Wafer Silicon WaferSiO2
Grow silicon dioxideSilicon Wafer
SiO2Resist
Apply photo resistSilicon Wafer
SiO2Resist
Mask Mask
Expose to UV
Silicon WaferSiO2
Patterned resistSilicon Wafer
Etch SiO2Silicon Wafer
Met
Deposit metalSilicon Wafer
Met
Etch SiO2(Or not)
![Page 30: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/30.jpg)
Building Blocks: Transistors
![Page 31: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/31.jpg)
Building Blocks: Wires
![Page 32: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/32.jpg)
State of the art CPU
• 1-2 Billion xtrs• 45nm features• 3-4Ghz• Several 100 designers• >5 years• $3Billion fab• 70 GFLOPS
18
![Page 33: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/33.jpg)
Current state of architecture
![Page 34: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/34.jpg)
Since 1940
![Page 35: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/35.jpg)
Since 1940
• Plug boards -> Java
• Hand assembling -> GCC
• No OS -> Windows Vista
![Page 36: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/36.jpg)
Since 1940
• Plug boards -> Java
• Hand assembling -> GCC
• No OS -> Windows Vista
Flexible performance is a liquid asset
• 50,000 x speedup
• >1,000,000,000 x density (Moore’s Law)
![Page 37: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/37.jpg)
Moore’s Law: Raw transistors
![Page 38: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/38.jpg)
The Importance of Architecture
• We design smarter and smarter processors
• Process technology gives us about 20% performance improvement per year
• Until 2004, performance grew at about 40% per year.
• The gap is due to architecture! (and compilers)
![Page 39: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/39.jpg)
Computer Performance
23
![Page 40: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/40.jpg)
Computer Performance
23
1
10
100
1000
10000
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Rela
tive
Perfo
rman
ce
Year
specINT95specINT2000specINT2006
![Page 41: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/41.jpg)
Computer Performance
23
1
10
100
1000
10000
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Rela
tive
Perfo
rman
ce
Year
specINT95specINT2000specINT2006
1
10
100
1000
10000
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Rela
tive
Perfo
rman
ce
Year
specINT95specINT2000specINT200647% per year
![Page 42: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/42.jpg)
Computer Performance
23
1
10
100
1000
10000
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Rela
tive
Perfo
rman
ce
Year
specINT95specINT2000specINT2006
1
10
100
1000
10000
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Rela
tive
Perfo
rman
ce
Year
specINT95specINT2000specINT200647% per year
1
10
100
1000
10000
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Rela
tive
Perfo
rman
ce
Year
specINT95specINT2000specINT200647% per year39% per year
![Page 43: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/43.jpg)
Computer Performance
23
1
10
100
1000
10000
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Rela
tive
Perfo
rman
ce
Year
specINT95specINT2000specINT2006
1
10
100
1000
10000
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Rela
tive
Perfo
rman
ce
Year
specINT95specINT2000specINT200647% per year
1
10
100
1000
10000
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Rela
tive
Perfo
rman
ce
Year
specINT95specINT2000specINT200647% per year39% per year
1
10
100
1000
10000
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Rela
tive
Perfo
rman
ce
Year
specINT95specINT2000specINT200647% per year39% per year25% per year
![Page 44: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/44.jpg)
The clock speed addiction
24
• Clock speed is the biggest contributor to power• Chip manufactures (Intel, esp.) pushed clock speeds very
hard in the 90s and early 2000s.• Doubling the clock speed increases power by 2-8x• Clock speed scaling is essentially finished.
0
1000
2000
3000
4000
5000
1996 1998 2000 2002 2004 2006 2008 2010
Cloc
k sp
eed
(Mhz
)
Year
specINT2000specINT2006
![Page 45: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/45.jpg)
Power
25
Watts/cm 2
1
10
100
1000
1.5µ 1µ 0.7µ 0.5µ 0.35µ 0.25µ 0.18µ 0.13µ 0.1µ 0.07µ
![Page 46: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/46.jpg)
Power
25
Watts/cm 2
1
10
100
1000
1.5µ 1µ 0.7µ 0.5µ 0.35µ 0.25µ 0.18µ 0.13µ 0.1µ 0.07µ
![Page 47: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/47.jpg)
Power
25
Watts/cm 2
1
10
100
1000
1.5µ 1µ 0.7µ 0.5µ 0.35µ 0.25µ 0.18µ 0.13µ 0.1µ 0.07µ
![Page 48: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/48.jpg)
Power
25
Watts/cm 2
1
10
100
1000
1.5µ 1µ 0.7µ 0.5µ 0.35µ 0.25µ 0.18µ 0.13µ 0.1µ 0.07µ
![Page 49: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/49.jpg)
Power
25
Watts/cm 2
1
10
100
1000
1.5µ 1µ 0.7µ 0.5µ 0.35µ 0.25µ 0.18µ 0.13µ 0.1µ 0.07µ
![Page 50: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/50.jpg)
Power
25
Watts/cm 2
1
10
100
1000
1.5µ 1µ 0.7µ 0.5µ 0.35µ 0.25µ 0.18µ 0.13µ 0.1µ 0.07µ
![Page 51: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/51.jpg)
What’s Next: Brainiacs
• Hold the clock rate steady.• Be smarter in silicon
• More sophisticated processors• More clever algorithms• This continues to deliver about 25% per year.• But for how long?
26
![Page 52: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/52.jpg)
What’s Next: Parallelism• This is all the rage right now• You probably own a multi-processor, they used to
be pretty exotic.• They provide some performance, but it’s hard to
use.• There aren’t that many threads• Remember, flexible performance is a liquid asset• Remember or look forward to cse121
27
![Page 53: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/53.jpg)
28
Intel P41 core
Intel Core 2 Duo2 cores
AMD Barcelona4 cores
SPARC T18 cores
Intel Prototype80 cores
Cell BE8 + 1 cores
Intel Nahalem4 cores
![Page 54: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/54.jpg)
Computer Performance
29
![Page 55: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/55.jpg)
Computer Performance
29
100
1000
10000
1996 1998 2000 2002 2004 2006 2008 2010
Rela
tive
Perfo
rman
ce
Year
specINT2000specINT200639% per year25% per year
![Page 56: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/56.jpg)
Course Staff
• Instructor: Steven Swanson• Lectures Tues + Thurs
• TA: Hung-Wei Tseng• Discussion sec: Wed. • (but not this week)
• See the course web page for contact information and office hours.
30
![Page 57: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/57.jpg)
What’s in this Class
31
• Course outline• Instruction sets• The basics of silicon technology• Measuring performance• How processors work
• Basic pipelining• Data and control hazards• Branch prediction and speculation
• The memory system• Introduction to multiprocessors
• Weekly technology digressions• How various technologies actually work.
![Page 58: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/58.jpg)
Your Tasks• Read the text!
• Computer Organization and Design: The Hardware/Software Interface (4th Edition) -- previous editions are not supported
• I’m not going to cover everything in class, but you are responsible for all the assigned text.
• Come to class!• I will cover things not in the book. You are responsible for
that too.• Class participation (5%)
• Homeworks throughout the course. (10%)• Weekly quizzes on Thursdays (10%)• One midterm. (25%)• One cumulative final. (35%)• One project (15%)
• Design your own ISA!32
![Page 59: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/59.jpg)
The Link to 141L
• You do not need to take 141L along with 141, but you may need both to get your degree.
• The classes are mostly independent, except• The results of the project will be used in 141L• You can earn extra credit by licensing your ISA groups in
141L who are not in 141
33
![Page 60: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/60.jpg)
Grading
• Grading is on a 13 point scale -- F through A+• You will get a letter grade on each assignment• Your final grade is the weighted average of the
assignment grades.
• An excel spreadsheet calculates your grades• We will post a sanitized version online once a week.• It will tell you exactly where you stand.• It specifies the curves used for each assignment etc.
• OpenOffice doesn’t run it properly.
34
![Page 61: 00 introduction](https://reader030.vdocument.in/reader030/viewer/2022033105/55d27f46bb61eb7a078b4601/html5/thumbnails/61.jpg)
Academic Honesty
• Don’t cheat.• Cheating on a test will get you an F in the class and no
option to drop, and a visit with your college dean.• Cheating on homeworks means you don’t have to turn
them in any more, but you don’t get points either. You will also take at least 25% penalty on the exam grades.
• Copying solutions of the internet or a solutions manual is cheating.
• Review the UCSD student handbook• When in doubt, ask. Honest mistakes will be
forgiven.
35