1. CIT 668: System Architecture
Computer Systems Architecture I

2. Topics System Components Processor Memory Storage Network
Operating System
Images courtesy of Majd F. Sakr or from Wikipedia unless otherwise noted.

3. A Single CPU Computer Components

4. Von Neumann Architecture
Single storage for both data and commands Storage separate from control unit
RAM
CPU
The Brain
von Neumann

5. The 5 Von Neumann Components
Input/Output
CPU and ALU
Memory

6. Motherboard

7. The Processor

8. The Processor
The Brain: a functional unit that interprets and carries out instructions (mathematical operations)
Also called a CPU (actually includes CPU + ALU)
Consists of hundreds of millions of transistors.

9. Instruction Set Architecture (ISA)
Every processor has its own set of operations
Written in binary machine language (ML)
Each instruction consists of operator + operand
Assembly language is humanly-readable ML
Category
Example
Arithmetic
Add, subtract, multiply, divide
Logic
And, or, not, exclusive ov
Program Control
Branch (conditional), call, jump
Data Movement
Move, load, store
I/O
Read, write

10. Source Code to Machine Language

11. Program Execution Cycle
Fetch/Execute Cycle
Fetch: retrieve instruction from memory
Execute: decode instruction, fetch operands, execute instruction, store results to memory (if any)
Fetch
Decode
Execute
Store

12. Synchronous Execution
Each step waits for click tick to begin
A 100 MHz processor uses a clock that ticks 100,000,000 times per second.
Fetch
Decode
Execute
Store
Cycle speed of 1 GHz:
billionths of a second

13. Processor Components

14. Processor Layout

15. From Sand to Silicon Wafers
Purification and heating to make a pure silicon crystal.
Sand, made up of 25% silicon, is the starting point
Slicing the crystal into
300mm wafers for use in photolitho- graphy.
100kg of % pure silicon crystal

16. Photolithography http://cnx.org/content/m1037/latest/

17. Packaging Wafer is cut into dies, each containing one processor.
Substrate, die, and heatspreader assembled to make final CPU product for use in a PC.

18. Processor Components: Control
Control Unit
Processor’s supervisor
Fetch/execute cycle
Retrieve instruction
Decode instruction
Execute actions
Program Counter (PC): stores address of instruction to be fetched.
Instruction Register (IR): has instruction most recently fetched.

19. Processor Components: Datapath
Register File: General purpose storage locations inside processor that hold addresses or values. Arithmetic Logic Unit: Set of functional units that perform arithmetic and logic operations.
A 32-bit processor has registers that are typically 32 bits in size.

20. Superscalar Architecture
Instead of one ALU, use multiple execution units
Some execution units are identical to others
Others are different: Integer, FPU, multi-media

21. AMD Athlon Architecture

22. Computer System Layers

23. Moore’s Law – Number of transistors doubles every 18 months
More transistors =
Cheaper CPUs
Higher speeds
More features
More cache
Bring 1ns (11.8 in) cable to illustrate why shrinking transistors can improve speed

24. Improvements in CPU Clock Speed

25. Computation Produces Heat
LN2 cooling

26. Processor Power Density
Sun’s
10000
Surface
Rocket
Nozzle
1000
Nuclear
Reactor
100
Power Density (W/cm2)
Pentium IV
Pentium III
Pentium II
8086 10
Hot Plate
Pentium Pro
4004
8008
8085
386
Pentium®
286
486
8080
Source: Intel 1
1970
1980
1990
2000
2010
Year

27. The Single CPU is History
What can you do with more transistors at same speed?
Multicore CPUs
Multiple processors (cores) on a single chip
Run different programs on each core
Some programs can be rewritten to run on multiple cores: PhotoShop, bzip2
Most of you are running multi-core
6-core opteron image from

28. 2011 Server CPUs Intel Xeon AMD Opteron IBM Power 7 Sun Niagara 3
Up to to 3.3 GHz
Up to 18 MB L3 cache
AMD Opteron
4, 6, 8, or to 3.2 GHz
Up to 12 MB L3 cache
IBM Power 7
4, 6, or to 4.25 GHz
4 MB L3 cache per core (up to 32MB for 8-core)
Sun Niagara 3
GHz
6 MB L2 cache

29. End of Moore’s Law Si ≈ 0.3 nm
Si ≈ 0.3 nm

30. Better Materials More Machines Better Algorithms
Where do we go from here?
Better Materials
GaAs
Diamond
More Machines
Cloud
Containers
Better Algorithms
Map Reduce
Fuzzy Logic

31. Processor-Memory Gap