Intro

Question 1

Define sequential computing.

Answer 1

• Executes a program that contains a discrete series of instructions.
• Instructions are executed one after the other. (one instruction is execute at a time)
• All instructions are executed on a single core/processor.

Question 2

Define parallel computing.

Answer 2

• A problem is broken into discrete parts that can be solved simultaneously.
• One part contains a set of instructions that are executed on a specific core/processor.
• Different parts are executed on different cores.

Question 3

1. What are situations where parallel processing will outperform serial computing?
2. Vice-versa?

Answer 3

1. Applying filters in image processing, matrix multiplication
2. serial I/O - file reading, recursive algorithms with strong dependencies.

Question 4

What are the requirements a program should meet to be prallelized?

Answer 4

1. Be able to be partitioned into discrete partitions that can be solved simultaneously.
2. Execute multiple instructions at the same time.
3. Spend less time using multiple computing resources than on a single resource.

Question 5

Define speedup

Answer 5

The ratio of time taken to solve a problem on a single processor, to the time taken on a parallel system.

Question 6

Define efficiency

Answer 6

The ratio of speedup to the number of processors. (how effectively the processors are utilized)

Question 7

Why are GPUs so good at graphics processing?

Answer 7

They have thousands of cores optimized for parallel processing. Graphics operations include manipulating pixels where the same operation is performed on a large number of pixels. These operations can be done simultaneously. GPUs make this process efficient by parallelizing this processes.