pre-requisite reading

• Sections 1.1 - 1.5 of Modern Operating Systems
• Introduction and Chapter 1 of The C Programming Language

Overview of modern computers

Most modern computers contain these basic essential parts:

• Processors

• Memory

• Disks

• I/O (devices that transfer from data to and from the computer)

• software programs are applications that run on the hardware they do not need to have full knowledge of the hardware in order to run

• this is because operating systems abstract any low level 1s and 0s of direct interaction with hardware and provide a nice layer that software can interact with

• we can divide certain programs into two ‘arenas’ of execution: ‘user space’ and ‘kernel’ space

• programs like browsers, file explorer and spotify run in user space, while programs that manage things like your internet connection and battery optimization run in kernel space.

• this is done so that the operating system maintains federation over what can access the hardware and when

• if a program requires access to the hardware e.g. to draw to the screen, output to speakers or to print a document. it can request a promotion to kernel space

CPUs

• At a low level, programs that run on a computer contain a series of specific instructions that the CPU can interpret, called assembly. when executing a series of instructions the CPU will fetch, decode and execute, in that order. the instructions are stored in highly and fast available storage known as registers

• these instructions can be executed in either user or kernel space.

• in kernel space the CPU can execute the whole the instruction space while user space it is restricted to only that which does not access the hardware directly

• in the history of CPU development, not too long ago CPU design looked somewhat like this

• this allowed the seperate cycles of execution to be done simultaneously.

• more modern CPUs will look something like this

• in this configuration multiple instructions can be fetched, decoded independently and simultaneously, stored temporarily, the picked up by a free execution unit

• CPUs for decades have been following a simple law, the number of transistors will double in an integrated, while cost stays similar. while this isn’t a real scientific law. it is something that we observe to be true and has held true for a very long time. This is known as moore's law. and was first observed in 1965.

• this is only starting to slow down know, as transistors are starting to reach the size where if we try to go any smaller, quantum effects start to appear

memory

memory is made up of different layers and is located at various points on a computer

when can roughly seperate the catagories of memory into a couple segments

Access Time	Name	Capacity	Location
1 nanosecond	registers	less then a kilobyte	CPU
2 nanoseconds	Cache	4mb	CPU/Motherboard
10 nanoseconds	Main Memory	1 - 32GBs	RAM
10 milliseconds	Disk storage	1 - however many terabytes you want	SATA/PCIE connections

each segment of memory has its own specialty and use pros and cons

for caches we can even divide this into different types

• L1 - inside the CPU very limited
• L2 - shared across cores/per core allocation