Ch. 8 - Memory Management

Question 1

What are the 3 most common forms of locality?

Answer 1

> Temporal Locality
Spatial Locality
Branch Locality

Question 2

Temporal Locality (Definition and Example)

Answer 2

A referenced memory location is likely to be referenced again soon after. (eg. loops)

Question 3

Spatial Locality (Definition and Example)

Answer 3

If a memory location is referenced, it is likely that nearby locations will be accessed soon after. (eg. array traversal)

Question 4

Branch Locality

Answer 4

Branching of code is limited to a few nearby possibilities.

Question 5

What are the benefits of using locality of reference?

Answer 5

> Reduce the number of page faults (access the same page over and over again)
More TLB and cache hits (less likely to go to disk to get data)

Question 6

What are some issues in sharing memory?

Answer 6

> Transparency (several processes need to co-exists, unaware of each other)
Safety (processes cannot corrupt each other or the OS)
Efficiency (low overhead needed)
Relocation (programs can run in different memory locations)

Question 7

What are the steps a program must go through before being executed by the CPU?

Answer 7

1. Compiling (generate object code)
2. Linking (combine object code into executable code)
3. Loading (copy the executable code into memory, do any necessary run-time linking with libraries)
4. Execution (dynamic memory allocation during program execution)

Question 8

Direct Placement (Memory Management Scheme)

Answer 8

Load entire user program (code, data, etc.) into the same memory location. The linker produces the same loading address for every user program.

Fastest memory management scheme

Question 9

Overlays (Memory Management Scheme)

Answer 9

Program is organized into a tree-like structure. Root of overlay always loaded into memory while sub-trees a re loaded as needed by overlaying existing code.

Question 10

Static Partitioning

Answer 10

Main memory is divided into a fixed number of fixed size partitions (small jobs, medium jobs, large jobs). Multiple jobs can run concurrently sharing the CPU (1 small, 1 medium, 1 large)

Question 11

Dynamic Partitioning

Answer 11

Allow for any number of programs to be loaded into memory as long as there is enough room. Each program is allocated the exact amount of memory it needs.

Question 12

What is an issue of using static partitioning?

Answer 12

Wasted memory if there isn’t a job to execute that can fit in each partition (internal fragmentation)

Question 13

What is an issue with dynamic partitioning?

Answer 13

When programs finish, the space freed may be too small to run a new program (external fragmentation)

Question 14

Internal Fragmentation

Answer 14

Waste of memory within a partition (difference between size of partition and size of process loaded)

Question 15

External Fragmentation

Answer 15

Waste of memory between partitions (scattered non-contiguous free space).

Question 16

Swapping

Answer 16

Move some or all of a process’ memory to disk, freeing up memory for use by the process or other processes.

Question 17

Why is swapping not desirable?

Answer 17

Moving process’ memory from memory to disk and vice versa is slow. Swapping is a necessary evil to prevent memory from becoming full.