Memory Management Questions and Answers

Answer: A

Correct answer: 15 bits

Selecting one of 32 segments needs \( \log_{2}32 = 5 \) bits, and an offset within a 1 KB segment needs \( \log_{2}1024 = 10 \) bits:

\[ 5 + 10 = 15\ \text{bits} \]

Answer: B

Correct answer: programmer

The size of an array is fixed by the programmer in its declaration, so that the translator can reserve the right amount of storage before the program runs.

Answer: C

Correct answer: A multi-homed firewall

A multi-homed firewall has an interface on each network it joins, so it both separates the subnets and applies rules to everything passing between them and the outside. A router or VLAN separates without filtering, and a modem does neither.

Answer: B

Correct answer: between 0 and segment limit

An offset is a distance from the start of the segment, so it is legal from 0 up to but not including the limit. Anything else refers outside the segment and is trapped.

Answer: B

Correct answer: program function

Values supplied to an expression so that a result can be computed are handled by functions, which take arguments and return a value into the expression.

Answer: D

Correct answer: The page table becomes large but internal fragmentation is low

A small page size reduces internal fragmentation, since the waste is on average half a page per process. But it multiplies the number of pages, so the page table grows - consuming memory and requiring more levels of indirection. It also reduces TLB coverage: each TLB entry maps less of the address space, so a given program touches more entries and the miss rate rises. And disk transfers become less efficient, because the fixed cost of a seek is amortised over less data. A large page size reverses all four: smaller table, better TLB coverage, more efficient transfers, but more internal fragmentation and more data transferred that may never be used. The compromise most systems have settled on is 4 KB, with support for huge pages of 2 MB or 1 GB where a database or a virtual machine host benefits from the reduced TLB pressure. Page size is always a power of two, so that splitting an address into page number and offset requires no arithmetic.

Answer: A

Correct answer: program counter

The program counter holds the address of the next instruction, and the fetch cycle begins by sending its contents to memory. The instruction register receives what comes back.

Answer: D

Correct answer: the page that has not been used for the longest time

LRU evicts the page whose most recent use lies furthest in the past, on the reasoning that a page unused for a long time is unlikely to be needed soon - the principle of temporal locality. It approximates the theoretical optimal algorithm, the third option, which evicts the page that will next be needed furthest in the future. That is unimplementable, since it requires knowledge of the future, but it is used as a benchmark against which real algorithms are measured. FIFO, the second option, evicts the oldest-loaded page and is simple but poor, and it suffers Belady’s anomaly: adding more frames can increase the number of page faults, which is deeply counter-intuitive. LRU belongs to the class of stack algorithms and cannot exhibit it. True LRU is expensive to implement, needing a timestamp or a stack updated on every reference, so real systems approximate it with reference bits - the clock or second-chance algorithm being the usual method.

Answer: D

Correct answer: global variable

A variable visible throughout the program and in all its routines is global. One confined to a single routine is local.

Answer: A

Correct answer: replacement algorithms that approximate LRU

The reference bit is set by hardware whenever a page is read or written, and cleared by the operating system. It is used by replacement algorithms that approximate LRU. The need arises because true LRU is too expensive. Maintaining an exact ordering of pages by last use requires a timestamp or a stack updated on every memory reference, which no practical hardware provides. The reference bit is set for free as a side effect of the access, and one bit is enough to distinguish pages used since the last sweep from pages not used at all. The second-chance or clock algorithm uses it directly: a page at the head of the queue with its bit set is spared and the bit cleared, so only a page unused since the last sweep is evicted. The additional-reference-bits refinement keeps a byte per page, shifting the reference bit in at intervals, which gives a coarse history rather than a single bit and approximates LRU more closely at modest cost.