Direct-Mapped Caches
What you'll learn: How direct-mapped caches use a simple modulo calculation to assign each memory block to exactly one cache location.
The Core Idea
In a direct-mapped cache, every block of main memory maps to exactly one specific location in the cache. There's no choice involved—given a memory address, you can immediately calculate which cache line will hold it using modulo arithmetic.
Think of it like assigned seating at a theater where your seat number is determined by a simple rule: seat = ticket_number % total_seats. If you have ticket #137 and there are 8 seats, you sit in seat 1 (137 % 8 = 1). Every time someone with ticket #137, #145, or #153 arrives, they all must use that same seat 1.
How It Works
Given a memory address and a cache with N lines:
- Calculate the cache index:
index = (block_address) % N - Store (or look for) that block at cache line
index
For example, with 8 cache lines:
- Memory block 0 → cache line 0
- Memory block 1 → cache line 1
- Memory block 8 → cache line 0 (wraps around)
- Memory block 9 → cache line 1
- Memory block 27 → cache line 3 (27 % 8 = 3)
The Trade-off
This scheme is extremely fast—just one modulo operation tells you exactly where to look. However, if your program frequently accesses blocks 0 and 8, they'll keep evicting each other from cache line 0, causing conflict misses even when other cache lines sit empty.
Key Takeaway: Direct-mapped caches use modulo arithmetic to assign each memory block to one fixed cache location, maximizing speed but risking conflicts when multiple blocks compete for the same spot.