CO1 WEEK 5 FACTORS AFFECTING PERFORMANCE OF CPU

Clock speed - N of cores - Cache size - RAM size - WORD size - Bus width

more cores usually faster processing - Higher complexity Improves performance tacks spit between cores

More than one processor =synchronised

small amount of memory which is part of CPU = holds instructions which likely to be reused - checks this before requesting data from RAM - saves fetching

L1 part of the CPU itself both being the smallest but fastest to access. Each level the cache gets larger but slower to access. More L2 and L3 memory available the faster the computer can run.

ROM - Non volatile - store a limited n of instructions - not changeable

Memory - stores data and instruction

Random Access Memory is used to temp store data and instruction being used by the current program running. The phrase ‘Random’ means that data and instruction can be read from or written to any part of the memory.

• word = number of bits which moves from one component to another • can only reference memory locations which fit within size of address bus • how many bits CPU can manipulate in one go

The amount of bits that can be transferred simultaneously from one device to another. Provides two way connection allowing data and instructions to travel on.

access data in memory one direction from CPU - address bus determines maximum amount of accessible locations

Address bus - Data bus - Control bus

two way to transmit commands,timing,specific status info

CPU is a sequential machine so carries one instruction at a time in order. To coordinate this it has a internal clock which runs at certain speeds which indicates how fast a CPU can run.

Measured in megahertz (MHz) or gigahertz (GHz) and corresponds to how many instruction cycles the CPU can deal with in a second. (A 4 GHz CPU performs four billion cycles a second.)

Where some people increase CPU clock speed to try to make their computer run faster. Possible to change max CPU clock speed in BIOS settings.

Parallel Processing where a task or problem is split up across all the available cores so each core tackle part of the overall problem. (In theory doubles or quadruples performance.) Multitasking where multiple tasks can be performed at the same time e.g. each core handles a seperate piece of software.

Doubling number of cores doesn’t necessarily double computer speed as this extra speed is used up when they communicate between each other.

More instructions/programs held in RAM and run at same time. Software responds faster. Can run more memory intensive software. Reduces use of Virtual memory. Less use of Hard Disk which slower to access.