EEL 3801

1
EEL 3801

• Part II
• System Architecture

2
Components

• Video Display Terminal self explanatory
• Keyboard self-explanatory
• Disk Drives self-explanatory
• System Unit contains the motherboard or the
  system board. Otherwise self-explanatory

3
Components (cont.)

• Random Access Memory (RAM) Electronic memory
  where the program and the data are kept while the
  program is running. It is volatile since the
  contents are lost if there is loss of power.
  Additionally, it is also called dynamic since its
  contents must be continuously refreshed.

4
Components (cont.)

• Read-Only Memory (ROM) BIOS Contains the
  information on the input output peripherals.
• CMOS RAM Keeps system setup information.
• Expansion slots Permit expansion of the system
  by adding special purpose boards such as modems,
  communication cards, etc.

5
Components (cont.)

• Power Supply Self-explanatory
• Parallel Port Output port that transfers a set
  of bits simultaneously. Typically used for
  printers. Allow for quick transfer of data but
  only for short distances.
• Serial port Output port where single bits are
  produced one by one. Slower, but useful for
  longer distances.

6
Components (cont.)

• Microprocessor Intel microprocessors are
  downwardly compatible with each othe.
• Programs written on older versions will run on
  the newer ones, but programs written for the
  newer versions will not run on the older ones.
• Read Section 2.1 of the textbook for more details.

7
System Architecture

• The Central Processing Unit (CPU) is the most
  important part of the computer. It consists of
  the Arithmetic logic Unit (ALU) and the Control
  Unit (CU).
• The ALU carries out arithmetic, logic and
  shifting operations.
• The CU fetches data and instructions and decodes
  addresses for the ALU.

8
System Architecture (cont.)

• Additionally, there may be a math coprocessor,
  which speeds up mathematical calculations, as
  well as many other support chips. However, they
  are all coordinated by the CPU.

9
The CPU

• The most basic tasks of the CPU are
• Find and load the next instruction from memory.
• Execute the instruction. This is composed of
  several sub-instructions that we will discuss
  later.

10
The CPU (cont.)

• The CPU, besides the ALU and CU, is composed of
  several other components
• Data bus Wires that move data within the CPU
  itself.
• Registers High-speed memory elements within the
  CPU itself on which can significantly speed up
  the performance of the computer.
• Clock A timing device whose ticks coordinate all
  individual operations that take place in the
  computer. These ticks are called machine cycles.

11
Registers

• Registers are special work areas inside the CPU
  that can store data and/or instructions.
• These memory elements are very fast.
• There are several registers on the Intel 8088
  family of microprocessors
• Data registers
• Segment registers
• Index registers
• Special registers
• Flag register

12
Data Registers

• Also called general purpose registers.
• Are used for arithmetic and data manipulation
  operations.
• Can be addressed as either 8 or 16 bit values, or
  as both.
• The 80386 and newer CPUs use 32-bit registers
  addressable as 16-bit ones.

13
Data Registers (cont.)

• There are several of these.
• The AX Register the accumulator register is used
  by the CPU for arithmetic operations.
• It is a 16-bit register, but can be addressed as
  two independent 8-bit registers called AH (for
  high) and AL (for low).

14
Data Registers (cont.)

• The BX Register the base register is also
  general purpose (like the AX).
• Has the ability to hold addresses for other
  variables (pointers).
• Also 16-bit that can be independently addressed
  as two 8-bit bytes (BH and BL).

15
Data Registers (cont.)

• The CX register the counter register best serves
  as the counter for repeating looping
  instructions.
• These instructions automatically repeat and
  decrement the CX register, and quit when it
  equals 0.
• Also 16-bit that can be independently addressed
  as two 8-bit bytes (CH and CL).

16
Data Registers (cont.)

• The DX Register the data register is also
  general purpose but has a special role when doing
  multiplication or division.

17
Segment Registers

• These registers are used to store memory
  locations of either instructions or data in main
  memory.
• These registers contain the base segment of the
  memory location where the memory segment begins.

18
Segment Registers (cont.)

• There are several of these
• The CS Register the code segment register
  contains the base location of the executable
  instructions that make up the program.
• Note that the base location can only address the
  initial location where these instructions can be
  found, not the entire segment..

19
Segment Registers (cont.)

• The DS Register the data segment register is the
  default base location in memory for variables
• The SS Register the stack segment register
  contains the base location of the run-time stack.
  
• The ES Register the extra register is an
  additional memory location where additional base
  locations can be stored.

20
Index Registers

• Contain the offset (the distance from the base
  segment) where a specific variable or instruction
  may be found. The base segment and the offset
  can uniquely identify any addressable location of
  any length in memory. Base segment offset
  memory location.

21
Index Registers (cont.)

• There are several of these
• The SI Register the source index takes name from
  the instruction used to move strings.
• SI usually contains an offset from the DS
  register, but can address any variable.

22
Index Registers (cont.)

• The DI Register generally acts as a destination
  for string movement instructions. Typically
  contains an offset for the ES register, but not
  necessarily so.
• The BP Register the base pointer register
  contains an offset from the stack register (SS).
  
• Used to locate variables in the stack.

23
Special Registers

• Do not fit into any other categories.
• The IP Register the instruction pointer register
  contains the offset of the next instruction to be
  executed.
• Combines with CS to form the complete address of
  the next executable instruction.

24
Special Registers (cont.)

• The SP Register the stack pointer register
  contains the offset from the beginning of the
  stack segment to the top of the stack.
• SS and SP combine to form the complete address
  for the top of the stack.

25
Flags Register

• One single 16-bit register whose individual bit
  positions serve as flags to indicate the status
  of the CPU or the result of some arithmetic
  operation.
• The individual positions are predefined, although
  not all 16 are defined.

26
Flags Register (cont.)

• Bit positions and flags
• 0 ? Carry flag Set when result of unsigned
  arithmetic operation is too large to fit into
  destination. Values are 1carry 0no carry.
• 1 ? undefined
• 2 ? Parity flag reflects the number of bits that
  are set in the result of an operation. Can be
  even or odd.

27
Flags Register (cont.)

• 3 ? undefined
• 4 ? Auxiliary carry set when operation causes a
  carry from bit 3 to bit 4. Rarely ever used.
• 5 ? undefined.
• 6 ? Zero flag Set when result of an operation
  results in zero. Used in jumping to other
  instructions based on comparison of two values.
  Has a value of 1when 0 0 when 0.

28
Flags Register (cont.)

• 7 ? Sign flag Set when result of an operation
  results in negative number. Value is 1 when
  negative 0 when positive.
• 8 ? Trap flag Determines whether or not the CPU
  will be halted after each instruction is
  executed. Allows Trace or stepping through a
  programs execution. Allows the programmer to
  control the CPU in this way through the INT 3
  instruction.

29
Flags Register (cont.)

• 9 ? Interrupt flag Makes it possible for
  external interrupts to occur. Interrupts can be
  disabled by setting this flag to 0. Controlled
  by the programmer through the CLI and STI
  instructions.
• A ? Direction flag controls the assumed
  direction used by the string processing
  instruction. Values are 1up 0down.
  Programmer can control this flag through the STD
  and CLD instructions.

30
Flags Register (cont.)

• B ? Overflow flag Like the Carry flag, but for
  signed arithmetic operations. Value is
  1overflow 0no overflow.
• C, D, E and F ? undefined

31
The Run-Time Stack

• The run-time stack is an important element in the
  execution of a stored program.
• It is a temporary holding area for addresses and
  data.
• It resides in the stack segment identified in the
  SS and SP registers.
• Each cell in the stack is 16 bits.

32
Run-Time Stack (cont.)

• The stack pointer holds the last element to be
  added or pushed into the stack.
• This is also the first element to be taken off
  the stack, or popped.
• This is referred to as Last-In-First-Out (LIFO).

33
The Run-Time Stack (cont.)

• There are three typical uses for the run-time
  stack
• If we want to save the contents of a register,
  the stack makes a great place to store their
  values temporarily.

34
The Run-Time Stack (cont.)

• When a subroutine is called from another part of
  the program, it is important that the processor
  return to the place where the function was called
  after it exits. The address of the instruction
  that called the subroutine is saved on the stack
  so as to be able to return to it later.

35
The Run-Time Stack (cont.)

• Local variables can be created when a subroutine
  is active and then popped off the stack when the
  subroutine returns to the calling instruction.
  This is done in an area inside the run-time stack
  called the stack frame.

36
The Run-Time Stack (cont.)

• Operations
• The push operation Used to put values of data or
  instructions onto the stack. There is only on
  place in the stack into which things can be
  inputted the top of the stack.
• mov ax,00A5 move 00A5 into AX
• push ax pushes content of ax into stack
• push bx assume BX has a value of 0001
• push cx assume cx has a value of 0002

37
The Run-Time Stack (cont.)

• The push instruction does not change the value of
  the source register (typically the ax register,
  but could be others). Rather it simply copies
  its value to the top of the stack.

38
The Run-Time Stack (cont.)

• The pop operation Used to remove the value in
  the stack pointed to by the stack pointer and
  places it in a register or memory location
  (variable). Immediately upon removing the
  element popped, the SP moves to the immediately
  previous element in the stack.
• pop ax pops stack and puts value into AX

39
The Run-Time Stack (cont.)

• Note that the value remains in the stack, but not
  being pointed by the stack pointer, it is subject
  to be overwritten by the next push operation.

40
Microinstructions

• Machine level instructions are not the lowest
  level instructions in the computer.
  Microinstructions are. These are very low-level
  operations that carry out the machine-level
  instructions.

41
Microinstructions (cont.)

• There are three basic ones
• fetch the control unit fetches the instruction,
  copies it into the CPU (register).
• decode this operation decodes the instruction as
  well as any operands specified by the
  instruction. If any operands, the control unit
  fetches the operand from main memory.

42
Microinstructions (cont.)

• execute the ALU executes the operation and
  passes the result operands to the CU, where they
  are returned to the registers and/or to main
  memory.
• Get next instruction
• Go back to step 1
• Microcode is the interface between the binary
  code level and the electronic level.

43
Memory organization of DOS

• The Intel 8086 processor can access 1 Mb of
  memory (actually, 1,048,576 bytes, which is FFFFF
  in a 20-bit address). This is called the Real
  Mode.
• The main memory is divided into RAM and ROM.
• RAM occupies low memory, and starts at 00000h and
  continues up to BFFFFh.

44
Memory organization of DOS (contd)

• ROM occupies high memory and begins at C0000 and
  continues to FFFFF.
• This is mostly used for the ROM BIOS (the hard
  disk controller).
• The BIOS contains diagnostic and configuration
  software, as well as input-output subroutines.

45
Memory organization of DOS (contd)

• Addresses begin with a hex address of 00000 and
  continue incrementally until FFFFF.
• DOS allows only the first 640kB of RAM to be used
  for programs.
• This is misleading because DOS (74kB) itself has
  to occupy this area as well.
• Remaining RAM used by video display and hard disk
  controller.

46
System Memory (cont.)

• The 80286 and more notably, the 80386 and 80486
  processors can run in Protected Mode.
• This means that they can radically increase the
  amount of memory they can address (16MB).

47
System Memory (cont.)

• The Pentium can address significantly more than
  that.
• Unfortunately, DOS can only run in real mode.
• However, Windows runs in protected mode and
  liberates the programmer from the 1MB memory
  limit.

48
System Memory (cont.)

• The 80386 and beyond processor also has the
  virtual 8086 mode, which allows concurrent real
  mode processes to be executed in by a single CPU.
• The total memory being used can total more than
  the available RAM. The processor uses external
  memory (hard disk drive or floppy) to page
  currently unused portions of the program to these
  devices.

49
Address Calculations

• An address is a number that refers to an 8-bit
  (byte) memory location.
• The addresses are numbered consecutively,
  starting at 00000h and going up to the highest
  location in memory, depending on the amount of
  memory available.

50
Address Calculations (cont.)

• Addresses can be expressed in one of two ways
• A 32-bit (16 16) segment-offset address. This
  combines a base location (the base segment) with
  the offset to represent the actual address. For
  example, 08F10100, where 08F1 is the base
  location (segment) from which to start counting,
  and 0100 is the offset, or how much to count.
  The address points to the first byte in the
  address.

51
Address Calculations (cont.)

• A 20 bit absolute address, which refers to an
  exact memory location. For example, F405Bh.
• Using 20 bits, the processor can only address 1
  Mb (actually, 1,048,576 bytes) of memory.

52
Address Calculations (cont.)

• But address registers are only 16 bits wide,
  limiting the addressable memory to 65,535.
• Thus, the segment-offset technique is used to
  expand the range of accessible memory beyond the
  65,535 limit.
• Thus, when addressing memory locations, the
  registers combine the values of two registers,
  the base segment and the offset.

53
Address Calculations (cont.)

• The CPU uses the segment and offset value to
  generate an absolute address. It adds the
  segment and the offset to create the absolute
  address. The segment value is always known to
  have an implied half-byte at the right (0000).
• Example Given an address such as 08F10100.
  The absolute address (20 bit) would be calculated
  as follows

54
Address Calculations (cont.)

• Segment value plus implied byte 0 8 F 1 0 h
• Add the offset value 0 1 0 0 h
• __________________________________________
• Absolute Address 0 9 0 1 0 h
• The advantages to the segment offset method is
  that it allows the program to be loaded into any
  segment address in memory without having to
  recalculate the addresses of all variables.

55
Address Calculations (contd)

• Furthermore, large data structures that occupy a
  large block of memory can be easily accessed by
  knowing their base segment and offset.