CacheSim

A simulator for modelize cache memory of CPU.

Getting Started

This project is for Computer Architecture course in AmirKabir University of Technology.

For input of this simulator we must provide these informations:

• Total cache size
• Block size
• Unified vs. split I- and D-caches (Von Neumann. vs. Harvard)
• Associativity
• Write back vs. write through
• Write allocate vs. write no-allocate

And also expecting these informations at the end:

• Number of instruction references
• Number of data references
• Number of instruction misses
• Number of data misses
• Number of words fetched from memory
• Number of words copied back to memory

Input Format

Important This simulator reads instructions from file

Here is the format of main cache attributes:

• "block size" - "unified or separated" - "associativity" - "write policy" - "write_miss policy"
  • Explanation for each of the entries:
  • "block size": Cache’s block size in bytes which should be a power of 2. (e.g. 8, 64, 1024...)
  • "unified or separated": An indicator of cache’s architecture. “0” Denotes Von Neumann and “1” denotes Harvard.
  • "associativity": An integer indicating how associative the cache is. Remember that the cache’s size (explained shortly) should be divisible by its associativity.
  • "write policy": A string that should be either “wb” (write-back) or “wt” (write-through).
  • "write_miss policy": A string that should be either “wa” (write-allocate) or “nw” (no-write allocate)
• Then, if is 0, your program should read a line as follows:
  • "unified size"
  • Which is the cache’s size in bytes while being a power of 2.
• Otherwise, if is 1, your program should read the following line:
  • "instruction cache size" - "data cache size"
  • Both of which are the respective caches’ sizes in bytes while being powers of 2.

Then in every line of input we face first Type of instructions and address of data or instruction in hex

For types we have this rules:

• 0 Data load reference
• 1 Data store reference
• 2 Instruction load reference

The number following the reference type is the byte address of the memory reference itself. This number is in hexadecimal format, and specifies a 32-bit byte address in the range 0-0xffffffff, inclusive.

Tip we use LRU replace policy for this cache simulator

Example

Example for input

16 - 1 - 1 - wb - wa
128 - 128
0 00000 data read miss (compulsory)
2 10000 instruction miss (compulsory, replaces 00000 if assoc=1 & unified)
2 20000 instruction miss (compulsory, replaces 00000 if assoc=2 & unified)
2 30000 instruction miss (compulsory, replaces 00000 if assoc=2 & unified)
2 40000 instruction miss (compulsory, replaces 00000 if assoc=4 & unified)
0 00000 data read miss (miss if assoc=1 & unified)
2 10001 instruction miss (miss if assoc=1 & unified)

Example for output

***CACHE SETTINGS***
Split I- D-cache
I-cache size: 128
D-cache size: 128
Associativity: 1
Block size: 16
Write policy: WRITE BACK
Allocation policy: WRITE ALLOCATE

***CACHE STATISTICS***
INSTRUCTIONS
accesses: 5
misses: 5
miss rate: 1.0000 (hit rate 0.0000)
replace: 4
DATA
accesses: 2
misses: 1
miss rate: 0.5000 (hit rate 0.5000)
replace: 0
TRAFFIC (in words)
demand fetch: 24
copies back: 0

Hope you enjoy