Computer Architecture

1
Computer Architecture

• What is computer architecture?

2
Computer Architecture

• What is computer architecture?
• 3 levels
• Organization, design, layout, and connectivity of
  computer components
• Also refers to organization of chips on the
  motherboard
• Logical organization of circuits on a
  microprocessor

3
Computer Architecture

• What is the difference between computer
  architecture and hardware?

4
Computer Architecture

• What is the difference between computer
  architecture and hardware?
• Hardware is the actual physical components of the
  computer
• Architecture is the design of the layout, and
  connectivity of these hardware components

5
Inside the System Unit

• Main components in the system unit

6
Inside the System Unit

• Main components in the system unit
• Motherboard
• Microprocessor chip or CPU
• Drives floppy, hard, ZIP, CD, DVD
• Memory RAM
• Power supply
• Wires and ribbon cables
• System bus
• Expansion slots and cards

7
Inside the System Unit

• How are these components connected?
• CPU, memory, system bus attached to mother board
• Drive controller cards attached to the system bus
• Controller cards attached to drives through flat
  ribbon cables
• Power supply connected to motherboard and all
  drives
• Expansion cards inside expansion slots that are
  attached to the mother board
• Memory and CPU attached through high-speed data
  bus (not through system bus)
• External expansion ports are provided through
  internal expansion cards attached to the system
  bus

8
Inside the System Unit
Ribbon Cables
Expansion Cards- Video- Sound- Network-
Modem- Ports
ZIP Drive
CD Drive
Data Bus
MMU
Floppy Drive
Hard Drive
9
The Motherboard

• Main components of the motherboard

10
The Motherboard

• Main components of the motherboard
• Microprocessor or CPU
• Memory RAM, ROM
• System Bus
• Drive Controllers
• Expansion slots and cards
• External ports
• Real-time Clock

11
CPU Central Processing Unit

• What is it?
• What does it do?

12
CPU Central Processing Unit

• What is it?
• A single IC called a microprocessor
• The brain of the computer
• What does it do?
• Performs arithmetic and logical operations
• Executes instructions to process data
• Receives instructions and data from RAM
• Processes the instructions
• Places results back in RAM to be displayed or
  stored

13
Memory

• What is memory?

14
Memory

• What is memory?
• Electronic circuitry that holds data and program
  instructions
• Often called the primary storage or store, which
  creates the confusion with disk storage
• Memory has direct link to the CPU
• Storage refers to disk storage communicating with
  the CPU over the system bus

15
Memory

• Types of memory?

16
Memory

• Types of memory?
• RAM
• Cache
• Virtual
• ROM
• CMOS

17
Memory RAM

• What is RAM?

18
Memory RAM

• What is RAM?
• Random-access memory
• Often called main memory
• Place where CPU temporarily holds data before and
  after it is processed

19
Memory RAM

• Properties
• Large grid of capacitors, each represents 1 bit
• Capacitors are grouped together into bytes
• Each byte has a memory address
• location of the byte in the memory grid
• Usually represented by a single index into a
  large 1D array
• Capacitors must have continuous power to maintain
  data integrity volatile!

20
Memory Cache

• Main memory technology is too slow for todays
  fastest chips!
• Faster memory chips are available
• Prohibitively expensive in main memory capacities
• Put smaller and faster memory between CPU and
  main memory to store frequently used information

MMU
System Bus
21
Memory Cache

• CPU looks in cache before going to main memory
• Much faster when info is found cache hit.
• Much slower if information not found in cache
  cache miss
• In practice, due to locality of memory
  references, entire memory system operates close
  to speed of cache memory!

Cache Memory
MMU
System Bus
22
Memory Cache
CPU
L1 Cache

• Even faster, but smaller, memory can be put
  directly on the chip called L1 cache.
• L2 cache sits between CPU and main memory over
  data bus
• CPU first looks in smallest L1 cache, then into
  L2 cache, and then finally out to main memory
• Called memory hierarchy.

L2 Cache
MMU
System Bus
23
Virtual Memory
CPU

• Often, main memory is not large enough to store
  all temporary values the CPU needs for processing
• Solution use hard disk space to create a large
  virtual memory space
• In fact, the entire addressable space!32 bit
  addresses, so 232 bytes of memory!4GB of memory
  possible!

L1 Cache
L2 Cache
MMU
Hard Drive
System Bus
24
Virtual Memory
CPU

• Hard drive added to memory hierarchy
• Main memory becomes cache for the hard drive.
• Hard drive is MUCH slower, but cache misses in
  main memory are much less frequent since large
  memory pages are brought in from disk each time a
  miss occurs

L1 Cache
L2 Cache
MMU
Hard Drive
System Bus
25
Virtual Memory
CPU
L1 Cache

• The difference in speed between the hard drive
  and main memory is so dramatic that often an
  additional cache is added in the hard drive
  controller!
• Yet another level in the memory hierarchy!

L2 Cache
MMU
Hard Drive
Disk Cache
System Bus
26
Virtual Memory
CPU

• The memory hierarchy is controlled by the OS and
  its operation is transparent to the programmer.
• Programmers just see an extremely large
  addressable memory!
• Virtual memory has other benefits as well.

L1 Cache
L2 Cache
MMU
Hard Drive
Disk Cache
System Bus
27
Virtual Memory
CPU

• Each application can have its own completely
  independent virtual memory!
• The OS will handle mapping the applications
  virtual memory to disk and the paging into
  physical main memory
• Programmer need not worry about conflicts

L1 Cache
L2 Cache
MMU
Hard Drive
Disk Cache
System Bus
28
Memory Hierarchy
CPU

• In summary, data access within the complete
  memory hierarchy proceeds as follows
• CPU requests a memory read
• CPU first checks L1 cache. If hit, stop!
• L2 cache is checked. If hit, stop!
• Main memory is then checked. If hit, stop!
• Worst case, memory manager makes memory page
  request from the hard disk controller
• Disk controller first checks its cache. If hit,
  stop!
• Finally, memory page is read from hard disk

L1 Cache
L2 Cache
MMU
Hard Drive
Disk Cache
System Bus
29
Memory Hierarchy
CPU

• Big questions
• How do we decide what to store in each cache when
  reading and writing to our virtual memory?
• How do we keep the data in the caches consistent?
• This is THE big problem with caches and there are
  many rules and theories.

L1 Cache
L2 Cache
MMU
Hard Drive
Disk Cache
System Bus
30
Memory ROM

• What is ROM?

31
Memory ROM

• What is ROM?
• Read-only memory
• Hard-wired memory containing instructions for
• starting the computer
• running system diagnostics
• and controlling low-level I/O (BIOS)
• Why not use RAM? Because RAM needs basic amount
  of control instructions to store data. Needs
  initialization.
• ROM BIOS (Basic Input Output System) provides
  basic level of control for low-level system
  operations such as disk access

32
Memory CMOS

• What is CMOS?
• Complementary Metal Oxide Semiconductor
• Semi-permanent memory
• Holds low-level system configuration information
• Disk capacities, disk configurations, disk
  formatting
• Very little power required to retain information
• Can be updated, but not good for frequent random
  access