SLR02 - Types of Processors

What is an instruction set?

Set of all instructions written in machine code that can be recognised + executed by CPU - CISC, RISC

CISC

• Complex Instruction Set Computers - Aims to complete task in as few lines as possible

• Could includes specific instruction for multiplying two numbers - example of complex instruction

• Less common today

RISC

• Reduced Instruction Set Computer - aims to use simple instructions that will be executed within single clock cycle

• Would separate complex instructions into simpler commands

Characteristics of CISC

• Used un laptops / desktop computers

• Complex hardware

• Large num complex instructions

• Multiple clock cycles per instruction

• Physically larger, more silicon, more expensive

• Greater energy consumption

• Can’t make use of pipelining

• Emphasis on hardware, compiler less work to do

• Smaller code size, high cycles per second

• Makes more efficient use of RAM than RISC

• Multiple addressing modes available

Characteristics of RISC

• Used in smartphones, tablets, embedded systems, popular by ARM processors

• Simpler hardware

• Smaller num simple simple, standardised instructions of fixed length

• Single machine cycles per instruction

• Smaller size, less complex, less silicone, cheaper produce

• Low energy required

• Can support pipelining

• Emphasis on software = compiler more work

• Larger code size, lower cycles per second

• Heavy use of RAM

• Fewer addressing made available

Co-processor

Any additional processor used for specialised task e.g. GPU

Difference between CPUs and GPUs?

• GPUs run slower

• CPUs good all rounder, GPUs highly specialised

• GPUs superior speed + efficiency only certain tasks

• CPUs excel at performing complex operations on smaller data sets

• GPU better simple operations on larger data sets

Why are GPUs good for rendering graphics?

• All segments of image assigned XYZ co-ordainates

• Light source assigned + given angle, brightness

• Once all data plugged in calculations carried out, include:

• 3D arithmetic using XYZ co-ordinates

• RGB calculation

• XYZ positions typically stored as floating point = lots of floating point arithmetic

• Once all calc done, they run through every pixel of finished image from cam perspective + generate 2D representation of 3D space

• Huge num calc done parallel

• Mass of calculations typically associated with graphic processing (Single Instruction Multilple Data) can take full advantage of parallel processing

• GPU able perform thousand floating point operations much quicker

Multicore processor

Single chip containing two or more independent processing units

CMPS

Chip multiprocessors - manufacturers integrate cores onto a single physical chip as a CPU - actually a CMP containing multiple CPU cores

Performance of a multicore processor is further enhanced by

• One chip shared cache

• Inter - core communication

Limitations of multicore

• Overheads involved with iner-core communications

• Some programs can’t make maximum use of all cores

Parallel processing

The processing of program instructions by dividing them between multiple processing or processor cores, while one processor fetching instruction, separate unit decoding previous, other unit executing

What limits parallel processing

• Task being carried out

• If software designed for parallel processing