1.1.2 - Types of processor Flashcards
What are RISC processors?
Their instruction set consists of simple, commands which take the same time to execute; one per system clock cycle. This means to carry out a basic task, several instructions are needed.
What are the advantages of RISC processors?
Pipelining is easier to implement, because they have a simple instruction design.
They require less circuitry, so are cheaper to design, give off less heat and are less power-hungry.
What are the disadvantages of RISC?
Low-level programs for RISC processors are longer and so take up more memory, which can cause bottlenecks.
What are CISC processors?
CISC is designed to execute complex, variable-length instructions, carrying out a number of low-level instructions as part of a single instruction.
What are the advantages of CISC processors?
Having more specialised instructions means that low-level programs for CISC processors take up less space in the main memory and are easier to translate to machine code.
What are the disadvantages of CISC processors?
It is much more difficult to apply pipelining with a CISC processor, meaning that these types of processors have traditionally been used for specialist computer systems.
Where are RISC processors commonly used?
RISC processors are widely used in mobile phones, tablet computers, and many small specialist devices, as the lower power requirements of RISC makes them far more suitable for battery-powered devices. Interestingly, RISC processors are also widely found in supercomputers, where the demand for high-speed parallel processing makes RISC a superior choice.
What is a GPU?
A computer running video or graphics intensive applications needs it’s own dedicated graphics processor. They are designed for parallel processing, and it is not unusual for them to contain 1,000+ cores. Each core performs a small part of the whole operation.
What are GPUs used for?
Although known for being used in gaming, they are also used for purposes like machine learning, during which thousands of cores are optimised to parallel process calculations (for use in neural network training algorithms).
