[Unit 1.1] Components of a computer Flashcards
Computer Systems
define main memory
storage locations that are directly accessible by CPU. faster but lower capacity than secondary storage
more expensive than secondary storage (per GB)
define secondary storage
storage the computer can store all data not used but needed in long term
what is RAM
Random access memory, where programs being run are temporarily stored:a
OS, programs, data running
Volatile (lost when power shuts off)
directly accessible by CPU
what is ROM
Read only memory.
can be read from but not written to
non volatile
smaller capacity than RAM
stores firmware/software rarely changed:
-BIOS, embedded systems.
what is flash memory
non volatile memory that can be changed to update BIOS
counts as ROM
what does the BIOS do
initialises hardware and boots up computer
loads OS from secondary storage to RAM
what is virtual memory
use of secondary storage as RAM by OS
how does virtual memory work
memory split by OS into pages
less used pages move into VM
when pages in VM are needed they can be swapped back into RAM
this is a slow process
what are the disadvantages of virtual memory
if it is used all the time the computer will pause when switching between apps.
this is called thrashing.
it lowers the life expectancy of SSD.
what is parallel processing
simultaneous processing of data
what are the four ways of achieving parallel processing
pipelining
GPUs
Multicores processing
distributed computing
how does pipelining achieve parallel processing. what are the disadvantages
Fetch decode and execute at the same time
increases instructions processed in given time
complicated to build
branching instructions degrade performance
how does multicore processing achieve parallel processing
2 or more independent processing units
each core executes instructions independently
-each core does its own pipelining
how does a GPU achieve parallel processing
it is a specialised processor to perform graphical calculations
allowing main CPU to perform other tasks
perform same calc to multiple data at same time
has thousands of cores
has its own ram
how does distributed computing achieve parallel processing
multiple computers work across a network to solve a common problem
what are the limitations of parallel processing
task interdependency (not all tasks can be parallelised)
scalability constraints
specialised software (multiple cores has to be coded to be used)
amdahls law (performance limited by parts of problem that cant be parallelised)
what are the advantages of parallel processing
faster for handling large amounts of data
not limited by von neumann bottleneck
is using maximum power of CPU
what is a CISC design for the CPU
Complex Instruction Set Computer
has a large instruction set built in
-hardware more complicated but software easier
single instruction may take many clock cycles
allows for powerful instructions at cost of slower and larger CPU
Desktops
what are the advantages of a CISC design for the CPU
simplifies programming
easier for compilers
easier for low level programmers
shorter programs
what are the disadvantages of a CISC design for the CPU
more hardware needed
more transistors
larger in size
more expensive
consumes more power
what is a RISC design for the CPU
Reduced Instruction Set Computer
has small highly optimised instruction set built in
-simpler hardware, harder software
complex instructions arent used frequently
each instruction carried out in 1 cycle of FDE
Phones and embedded systems
what are the advantages of a RISC design for a CPU
less transistors
easier to design
smaller in size
more power efficient
cheaper
pipelining can be implemented easier
what are the disadvantages of a RISC design for a CPU
more software for complex tasks
longer programs
large number of general purpose registers needed
what other uses do GPUs have
cryptomining
AI and ML
scientific simulations
modelling
editing
gaming
data analysis
cryptography
VR and AR