Paging trade-offs: Which of the following statements about page size and paging performance is FALSE?

Introduction / Context:
Choosing a page size affects memory overhead, fragmentation, locality, and I/O efficiency. Understanding these trade-offs helps in OS and architecture design and in tuning applications for systems with different page sizes.

Given Data / Assumptions:

• Internal fragmentation occurs when allocated space exceeds requested space inside a fixed-size unit.
• Page table size grows with the number of pages.
• Larger I/O blocks can amortize overhead, often improving throughput.

Concept / Approach:

Smaller pages reduce internal fragmentation because wasted space per allocated page decreases. However, smaller pages increase the number of pages needed to cover an address space, increasing page table size and TLB pressure. Bigger pages improve sequential disk I/O efficiency but may bring in unnecessary data and instructions, potentially increasing cache and memory bandwidth waste.

Step-by-Step Reasoning:

Verification / Alternative check:

OS literature and performance studies confirm the general relationships: small pages reduce internal waste but increase metadata overhead; large pages improve I/O and TLB reach but can cause overfetch.

Why Other Options Are Wrong:

• A: Correct statement—page tables grow with smaller pages.
• C: Correct—overfetch happens with large pages.
• D: Correct—larger, aligned I/O transfers are typically more efficient.
• None of the above: Incorrect since B is false.

Common Pitfalls:

Confusing internal with external fragmentation; assuming larger pages always improve performance (they may hurt locality for sparse accesses).

Final Answer:

internal fragmentation is increased with small pages