Cheng-Chang Yang
 Generally speaking, faster memory is more
expensive than slower memory.
 To provide the best performance at the reasonable
cost, memory is organized in a hierarchical
function.
Memory Hierarchy
 The base types that
hierarchical
memory system
include:
 Registers
 Cache Memory
 Main Memory
 Secondary Memory:
hard disk, CD…
What is Cache Memory?
 Cache memory is to speed up memory accesses by
storing recently used data closer to the CPU,
instead of storing it in main memory.
Cache and Main Memory
Cache and Main Memory
Cache and Main Memory
 The Level 2 cache is slower and larger than the
Level 1 cache, and the Level 3 cache is slower and
Larger than the Level 2 cache.
 Transmission speed
 Level 1 > Level 2 > Level 3
 Transmission capacity
 Level 1 < Level 2 < Level 3
Flow Chart (Cache Read Operation)
RA: read address
Cache Mapping Function
 How to determining which main memory block
currently holds a cache line?
 Direct Mapping
 Associate Mapping
 Set Associate Mapping
Direct Mapped Cache

Direct Mapped Cache
 Maps each block of main memory into only one
possible cache line.
Associative Mapping
 Instead of placing main memory blocks in specific
cache locations based on main memory address, we
could allow a block to go anywhere in cache.
 In this way, cache would have to fill up before any
blocks move out.
 This is how associative mapping cache works.
Associative Mapping
 Associative mapping overcomes the disadvantages of
direct mapping by permitting each main memory block
to be loaded into any line of the cache.
Associative Mapping
 We must determine which block to move out from the
cache .
 A simple first-in first-out (FIFO) algorithm would
work. However, there are many replacement
algorithms that can be used; these are discussed in
later.
Set Associative Mapping
 The problem of the direct mapping is eased by having
a few choices for block placement.
 At the same time, the hardware cost is reduced by
decreasing the size of the associative mapping search.
 Set associative mapping is a compromise that exhibits
both the direct mapping and associative mapping
while reducing their disadvantages.
Replacement Algorithms
 For direct mapping there is only one possible line for
any block, and no choice is possible.
 For associative and set associative mapping, a
replacement algorithms is needed.
 Least recently used (LRU)
 First in first out (FIFO)
 Least frequently used (LFU)
 Random
Replacement Algorithms
 Least recently used (LRU) algorithm keeps track of the
last time that a block was assessed and evicts the block
that has been unused for the longest period of time.
 First in first out (FIFO) algorithm: the block that has
been in cache the longest would be selected and
removed from cache memory.
Replacement Algorithms
 Least frequently used (LFU) algorithm: replace that
block in the set that has experienced the fewest
references.
 The most effective is least recently used (LRU)
Reference
 Internet Source
 Wikipedia
(http://en.wikipedia.org/wiki/Cache_memory)
 Book
 Computer Organization And Embedded Systems(6th)
 Computer Organization And Architecture(8th)