Explore why paging and segmentation solve memory management challenges in operating systems. Compare fixed and dynamic partitions, internal and external fragmentation, and how paging enables non-contiguous allocation.

Revisit internal fragmentation in paging, showing it can occur when the process size is not a multiple of the page size or when the page table is smaller than a page, while paging avoids external fragmentation.

Compute ram size as 2^(m-n) frames times 2^n bytes per frame, giving 2^m bytes. Relate processes to paging: 2^l pages with 2^k bytes per page yield total size 2^(l+k) bytes.

Introduce multilevel paging to handle non-contiguous process pages by using level 0 and level 1 page tables stored in RAM frames, with pointers, simple indexing, and awareness of internal fragmentation.

The lecture analyzes a three-level paging system, derives 2^10 outer page table entries, 2^20 level-two page table entries, and 2^30 level-one entries, yielding an 8 tb address space.

solve a single-level paging problem by computing 2^10 page table entries from an 8 KB outer page table with 8-byte entries, yielding a process size of 2^23 bytes.

The dirty bit in a page table entry signals whether a page was modified and must be written back during page replacement, avoiding unnecessary writes to disk and saving time.

Compute the page table size in a 32-bit, byte-addressable system with 4 KB pages and 4-byte entries, showing that 2^20 entries yield 2^22 bytes or 4 MB.

Explain locality of reference and virtual memory, showing how the cpu uses only a few pages at a time to boost degree of multiprogramming, via ram frames and disk storage.

Explore average memory access time for a two-level paging system with 10 ns ram and 1000 ns disk, 90% page hit rate, yielding 1030 ns and 130 ns.