The Von Neumann Model

1
The Von Neumann Model
2
The whole model

• Four basic components
• Memory
• A processing Unit
• Input
• Output
• Control

MDR Memory Data Register MAR Memory Address
Register PC Program Counter IR Instruction
Register ALU Arithmetic and Logic Unit LED
Light Emitting Diode
3
Something about memory

• Memory in its most general sense refers to places
  to put things for later retrieval.
• Some special memory for temporary storage
  (registers) is located on the processor chip.
• Access to this memory is extremely fast, but it
  is very expensive and in very small supply (16,
  32, that sort of number).
• Mostly, when we talk about memory, we mean RAM --
  Random Access Memory

4
RAM

• An array of bits
• Addressability refers to how many bits are
  associated with each address provided.
• N bit address can access 2N locations
• In our 4 x 3 example, there were four addresses,
  each of which specified a set of three bits
• Power required to store the value
• Gated D-latch type of circuit
• Has to be loaded after the computer is powered up

5
RAM continued

• Amount of RAM is now measured in Gigabytes
• Giga means billions
• Kilo (K) is actually 1024
• Mega (M) is actually 1,048,576
• Giga (G) is actually 1,073,741,824
• byte means 8 bits (almost always)
• How big an address would be required to access
  each byte in an 80 GB memory?
• Word size is the number of bits accessed at one
  time in a particular machines architecture

If the computers of today contain 2 to the 28th
by 8 bits, then how does the 256 megabyte memory
come into play? What exactly is 256?
6
More RAM

• To access the content of a memory location,
  specify the location (its address)
• Location and content are two distinct, very
  important concepts.
• We will see the distinction clearly when we study
  the C language. (C allows us to manipulate the
  address, as well as the content)
• Memory can be read from or written to.
• Reading is non destructive -- what was there
  before remains
• Writing is destructive -- what was there before
  is replaced by the new information.

7
Reading and Writing RAM

• MAR contains the address to be used (the values
  on the A lines in our map of a small memory)
• MDR contains the data to be written (if writing)
  or will receive the value read (if reading)
• These are two registers (located on the processor
  chip) that are dedicated to these specific
  purposes.

8
Processing Unit

• The location of all the circuitry related to
  manipulating the bits for achieving some purpose
• The Arithmetic and Logic Unit
• performs arithmetic and logical operations on
  data stored in registers, usually. (It would be
  very inefficient to have these circuits try to
  access RAM directly)
• Registers
• local storage for data in immediate use

9
Control

• We have memory to store instructions and data,
  and the processing unit to carry out operations.
• The control unit determines what happens next and
  keeps the processing unit supplied with the next
  instruction.
• Good question asked How are these instructions
  which allow the computer to operate defined.
  Meaning were do they originate from (RAM, bios,
  etc.) and how do they function in the processor.

10
From the beginning
R0 R1 R2
IR
PC
MAR
The bare machine without power. Nothing is
happening.
MDR
The control unit controls the rest of the
computers functions. How does the control unit
figure out what it has to do first and where it
looks for the components it needs to run.
Memory
- empty
11
To get started
Something else is needed. ROM stores the
instructions needed to get started. When power
is available, the circuitry in the computer loads
the contents of ROM into a predetermined location
in memory.
R0 R1 R2
IR
PC
MAR
MDR
ROM
Memory
-
12
First steps
R0 R1 R2
IR
PC
MAR
Now, these few instructions control the computer.
They have one purpose to load the operating
system core into memory and turn over control of
the computer to the operating system.
MDR
00..0 00010100 00..1 00..2 00..3 Etc.
Memory
-
13
Under control of the OS
R0 R1 R2
IR
Now the OS is in control. The user commands to
the operating system asks for the appropriate
application program to be loaded. In turn, the
application program calls for its data as needed.
PC
MAR
MDR
Memory
- OS loaded
14
What operations?

• Thats what makes machines different
• The list of things that a given machine can do is
  called its Instruction Set
• The Instruction Set Architecture (ISA) is the
  description of what a given machine can do,
  including the instructions it can execute and the
  amount of local memory it has available to work
  with.
• An instruction is a combination of bits with a
  specific meaning.
• Think of an instruction as a set of switch
  settings

15
Instructions

• Can you wash your clothes in your microwave oven?
• Of course not, because that is not in the ISA of
  a microwave.
• Are all microwave ovens alike?
• No. They vary in power and in conveniences such
  as rotating trays.
• Some microwave ovens have an instruction (a
  setting) for baking a potato. Others do not have
  that instruction, but can do the same thing if
  you give it the right steps to follow.

16
CPU

• The central processing unit of the computer is
  where the work gets done.
• The instruction set of a particular computer
  consists of the recognized instructions -- ADD,
  AND, NOT, OR, Subtract, Multiply, Divide, Move,
  Compare, etc.
• The list varies from one machine to another.
• All general purpose computers are equivalent.
  Given enough memory and enough time they can do
  the same things. (They are all Turing machines.)

17
Input and Output

• A computer is not very useful if we are unable to
  give it commands and see what it has done
• Input is a way of getting instructions and data
  into the computer.
• Mouse, Keyboard, sensors, scanners,
• Output is a way of getting results out of the
  computer
• Monitor, printer, plotter, etc.
• Some devices are used for both input and output
• CD-RWs, various disks, network connections, etc.

18
The instruction cycle

• Basic version fetch and execute
• More detailed version
• fetch, decode, evaluate address, fetch operands,
  execute, store result
• Underlying theme
• Instructions and data are stored in RAM until
  needed.
• To execute an instruction, the CPU has to find
  the instruction, move it into the IR, find the
  data, move it into registers, carry out the
  instruction and store the result back in memory

19
The LC-2
In Fig 4.3, what are the 3 arrows marked DR, SR1,
and SR2 that go out of the control logic box?
Where do they output to?
DR Destination register SR Source
Register Any others?
the LC-2 drawing on pg 80 confuses me, because
the PC seems not to connect to the MAR or main
memory or anything, and i dont know what the
purpose of those two big arrows pointing to the
left are.
20
Finding the instruction and data

• The MAR contains the address of the instruction
  or the data to be fetched. How the MAR gets its
  value is an important part of how the computer
  works.
• In the normal cycle, the PC is incremented to
  point to the next physical location. (Not all
  computers require that the PC contents be moved
  to the MAR. The PC can serve as an MAR for
  fetching instructions.)
• To fetch data, the MAR must be loaded explicitly
  from information provided in the instruction.

21
Example instructions

• In each type of computer, there is a specific
  format for each type of instruction. (Refer to
  the IBM 360/370 ISA distributed earlier.)
• Each bit position has meaning
• A bit is part of the operation to be performed or
  part of the specification of the location of data
  or the location where the result will be stored.
• There are a number of formats for specifying the
  location of data. More about that soon.

22
ADD in the LC-2

• The LC-2 ADD instruction
• Bit pattern 0001 means ADD
• The source and destination addresses are
  registers
• Bit positions 53 are not used
• What does instruction
• 0001010101000111 mean?

0
15
Result
Source 1
Source 2
Op Code
How do you know that something is in an ADD
instruction format and not an LDR instruction
format? Does it have to be labeled? And how would
you do an instruction format?
23
The cycle

• How do you know that something is in an ADD
  instruction format and not an LDR instruction
  format? Does it have to be labeled? And how would
  you do an instruction format?
• This is why we have several steps to the fetch
  execute cycle. After fetching the instruction,
  we then decode it. Once we know which operation
  we have, we know what format the addresses will
  be in. Then we can evaluate the addresses and
  then we can fetch the operands.
• Only then can we execute the instruction and
  store the results.

24
Other formats

• The LDR example

15
0
0
0
1
1
0
0
1
0
0
1
1
0
1
1
1
0
DR
LDR
Base register offset amount
This is just a sample. We will see other
addressing formats in the next chapter.
25
Changing the flow

• JMPR
• The JMP is a Jump instruction and will specify
  exactly where the next instruction is to come
  from.
• The R at the end says that the addressing format
  used is base displacement
• The address obtained will be written into the PC
  to control where the next fetch cycle will go to
  find an instruction.

26
The clock

• As we saw with the simulator, a clock in a
  computer is just a pulse generator. It cycles at
  regular intervals and controls the events
  occurring. When the clock goes faster, everything
  in the machine goes faster. If the clock goes
  too fast, the operations will not be finished and
  the state of the machine will be confused. If
  the clock goes too slow, the machine will run
  slower than it needs to.