Virtual Memory

• CPU and OS translate virtual addresses to physical addresses
• Initially designed so you don’t need to store your program and data separately
• Now it is used for protection
  • Allows the OS to check that you have access to the certain physical address
  • For each address check base ⇐ address ⇐ bound

Four Important Questions

• Where to place a block in main memory?
  • Operating system takes care of it
  • Replacement takes very long because it is fully associative
• How to find a block in main memory?
  • Page table is used
  • Offset is concatenated when paging is used
  • Offset is added when segmentation is used
• Which block to replace when needed?
  • LRU is used to minimize page faults
• What happens on write?
  • Always write back (use of dirty bit)
  • Because magnetic disks takes millions of cycles to access

Fast Address Calculation

• Page tables are very large
  • Kept in main memory
• Remember the last translation
  • Reuse if the address is on the same page
• Exploit the principle of locality
  • If access have locality, the address translations also have locality
  • Keep the address translations in a cache
    • Translation lookaside buffer (TLB)
    • The tag part stores the virtual address and the data part stores the page number.

Addressing

• We will cover paging
  • Intel uses segmenting for backwards compatibility
• We have a table that translates between the virtual address and the paged address
  • This table grows as needed, because the whole table would be petabytes
    • It is also stored across levels of caches
    • Translation lookaside buffer
  • Along with the optimizations
    • Store the virtual address implicitly using index
    • Use inverted table (physical to virtual) because there are less physical addresses
• While each process feels like it has access to all the memory, it is expected for them to not abuse that