virtual memory

Concept

• secondary storage is usually larger than memory
• extension of the paging scheme, some pages can be stores on the secondary storage
• OS required to manage the virtual memory
• like a second layer of caching

Page fault

• pages can only be accessed when in memory, CPU cannot read data from secondary storage
• when CPU tries to access a non-memory resident page
• triggers the OS to bring the page into physical memory

Thrasing

• if memory access results in page faults most of the time
• accessing secondary storage is very slow, so this can slow down the system

Locality

• larger page size -> exploit spatial locality

Demand paging

• process starts with no memory resident pages
• pages are only allocated when there is a page fault
• pros:
  • fast startup time
  • small memory footprint
• cons:
  • process may be sluggish at the start
  • page faults may have cascading effect on other processes, thrashing