The CPU - Chapter 1

Question 1

What is the ALU?

Answer 1

Arithmetic Logic Unit

Question 2

What is the purpose of the ALU?

Answer 2

The ALU carries out calculations and logic decisions.

Question 3

What is the Control Unit?

Answer 3

The Control Unit co-ordinates signals across the processor, and controls how the data moves around the CPU and between memory

Question 4

What is a PC?

Answer 4

Program Counter - keeps track of the line of machine code that is being worked on. It increments every cycle.

Question 5

What is the ACC?

Answer 5

The Accumulator stores the results of any immediate calculations

Question 6

What is the MDR?

Answer 6

Memory Data Register stores data that has been fetched from memory

Question 7

What is the MAR?

Answer 7

The Memory Address Register stores the address of the data that has been fetched from memory

Question 8

What is the CIR?

Answer 8

The Current Instruction Register stores the most recently fetched instruction waiting to be decoded.

Question 9

How much L1 Cache on each core?

Answer 9

2 lots

Question 10

How much L2 Cache on each core?

Answer 10

1, it is slower but more storage than L1

Question 11

How much L3 Caches on each core?

Answer 11

Trick question - there is 1 L3 Cache allocated to the entire CPU. It has a lot more storage but as a result it is much slower.

Question 12

What is a Micro Processor?

Answer 12

A dedicated processor for a limited purpose - such as a washing machine.

Question 13

What are the two main limitations of a CPU?

Answer 13

Clock Speed and Core count.

Question 14

What of the two main limitations of the CPU has a limit? Why?

Answer 14

Clock Speed can only go so high, before tunneling occurs. This means clocks are skipped.

Question 15

What is SoC?

Answer 15

System on a Chip is an integrated circuit that integrates all components on a system, such as a Raspberry Pi

Question 16

What is RISC?

Answer 16

Reduced Instruction Set Computer

Question 17

What is CISC?

Answer 17

Complex Instruction Set Computer

Question 18

Properties of CISC?

Answer 18

Larger set of instructions
Requires more complex hardware
May take more cycles per line
Greater Energy consumption
Cannot support pipelining
Used in large machines such as desktop PCs

Question 19

Properties of RISC?

Answer 19

Limited set of instructions
Requires more lines of code for simple operations
One cycle per line
Little energy consumption
Can support pipelining
Used in small devices such as mobile phones

Question 20

What is pipelining?

Answer 20

Pipelining is the process of passing any complete process onto the next step of the cycle

Question 21

When will the CPU ‘Flush the pipes’?

Answer 21

When an instruction that is further away than the next (such as branching to instructions further ahead or behind)

Question 22

What are the two types of Multi-core systems?

Answer 22

Parallel systems and Co-Processors.

Question 23

What is the difference between Serial computing and Parallel computing?

Answer 23

Serial computing does everything in one thread, one after the other.
Parallel computing can get through multiple problems at once - ‘in parallel’.

Question 24

Why must the split up problems be discrete?

Answer 24

We do not split up the same problem into different cores. This is because all processors will have to wait for the others to finish their task – eventually the task will have to be pieced back up together at the end.

Question 25

What is SIMD? What is it used for?

Answer 25

Single Instruction, Multiple Data. It is useful for image processing, or any task with a high degree of regularity.

Question 26

What is MISD? What is it used for?

Answer 26

Multiple Instructions, Single Data. It is not a common type of parallel system, but it has its uses in very specialized decryption.

Question 27

What is MIMD? What is it used for?

Answer 27

Multiple Instructions, Multiple Data. Most modern computers and supercomputers fit into this category. It is good for multitasking with different programs and data.

Question 28

Why did programmers used to avoid Parallel systems?

Answer 28

It was very complicated to join all the threads together at the end. It used to be quicker to have a single core computer as a result. However as programmers have become more adept to writing code for parallel systems, these original issues have become less of a problem today.