1.1.2 Types of processor Flashcards
What does “CISC” stand for?
Complex Instruction Set Computing.
What does “RISC” stand for?
Reduced Instruction Set Computing.
What are some of the differences between CISC and RISC processors?
RISC have a smaller range of instructions that CISC.
As RISC have smaller instruction sets, less transistors are needed meaning less power is required, thus making RISC processors cheaper to make than CISC.
CISC instructions take multiple machine cycles to execute, where as RISC instructions only take a single cycle, meaning they can benefit from Pipelining.
Due to the complexity of RISC Compilers, more instructions are generated for programs compared to CISC.
RISC Processors have fewer addressing modes that CISC, but have more General Purpose Registers.
Where are “CISC” Processors used?
CISC Processors are usually made by Intel and AMD, and are generally used in Computers and Laptops.
Where are “RISC” Processors used?
RISC Processors are usually made by ARM, and are generally used in portable devices such as Smartphones and Tablets.
What does “GPU” stand for, and what process are they specialised to do?
“Graphical Processing Units” are specialised in rendering graphics.
Why are GPUs specialised in rendering graphics?
Because they can perform the same calculation on multiple pieces of data at very high speeds in parallel, often referred to SIMD.
What other fields are GPUs being applied to?
Modelling Physical Systems, Audio Processing, Breaking Passwords and Machine Learning.
What is “Parallel Processing”?
When a computer carries out multiple computations simultaneously to solve a given problem.
What does “SIMD” stand for, and what does it do?
“Single Instruction, Multiple Data” is an approach of Parallel Processing. It carries out a single instruction on multiple sets of data at once, and is generally used by GPUs.
What does “MIMD” stand for, and what does it do?
“Multiple Instructions, Multiple Data” is an approach of Parallel Processing. It carries out multiple different instructions on multiple sets of data concurrently. This can be carried out using multiple cores on a single CPU, or using thousands of CPUs across a super computer.
What is a drawback of using “Parallel Processing”?
Not every problem is parallelisable, meaning Parallel Processing cannot always be used to improve the time it takes to execute a problem.