That Blue Square Thing

AQA Computer Science GCSE

September 2020: This page is the for the 2022 exam version of the AQA Computer Science GCSE course. This is for Year 10 teaching in September 2020.
These pages were copied over from the old syllabus and will get updated as I work through the new one. That means they won't always be totally up to date and I might miss something - check the updated date at the bottom of the page.

thumbnail of COU and cache diagram

Computer Systems - CPU Performance

What makes a CPU perform better?

PDF iconCPU Performance Summary - slides from class

PDF iconCPU Performance - "textbook" notes on the factors

PDF iconComparing CPUs - screenshots to compare the machine I need to buy next...

The balance between the amount of main memory, caches and the clock speed and number of cores in a machine is one that needs to be carefully considered. And that's before other factors come into play - portability, operating system, software that will run on it, monitor size etc...

For example, increasing just the cache size won’t mean that the machine necessarily performs better - it depends on what the machine is being required to do and what other resources are available.

Small, ultra-portable laptop machines tend to have a lower technical specification because of the ways they are usually used. This might be absolutely fine - the tradeoff between portability and power is a sensible one to make at times.

Another obvious tradeoff is between cost and performance.


Two of the four factors relate to caches. These are memory areas closer to the CPU than main memory which provide temporary storage. They can be used to store data and instructions the CPU has used recently.

Caches are quicker areas of memory, often with a quicker bus. This means the data and instructions stored in them can be retrieved more quickly, improving performance.

The bigger the cache, the more can be stored in it - but caches are expensive because they're quicker memory, so there needs to be a tradeoff.

The process works something like this:

  1. data or instruction used by CPU
  2. dump it on the top of the cache
  3. if the cache is full the data at the bottom of the cache falls out

Then the next time the CPU needs to retrieve data or instructions, it looks in the cache for it first. If it's in the cache, it retrieves it and then dumps it back on the top of the cache. This means that stuff that the CPU is using lots stays at the top of the cache and never falls out - so it's always quicker to retrieve it.

There are three levels of cache - L1, L2 and L3.

Diagram from my board in class showing relationship between the CPU, caches, main memory and the harddrive.