Skip to content

A SystemVerilog-based simulation and design of a Last Level Cache (LLC) implementing the MESI protocol, featuring Pseudo-LRU replacement, multi-mode simulations, and comprehensive testing strategies for optimized processor performance.

Notifications You must be signed in to change notification settings

Sidshx/LLC-cache-simulator

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 

Repository files navigation

Last Level Cache Design and Simulation

This project focuses on designing and simulating a Last Level Cache (LLC) for processors, addressing the challenge of memory systems lagging behind high-speed processors. By implementing a caching mechanism, critical data and instructions are stored closer to the processor for faster access. This project is fully parameterizable and can be easily adapted to changes in the given specifications.

Key Features

L2 Cache Specifications

  • Total Capacity: 16MB
  • Byte Line: 64 bytes
  • Associativity: 16-way set
  • Coherence Protocol: MESI
  • Replacement Policy: Pseudo-LRU

L1 Cache Specifications

  • Byte Line: 64 bytes
  • Associativity: 4-way set
  • Write Policy: Write-once (initial write-through, subsequent write-back)

Design Architecture

image

Functional Highlights

  • Implements MESI protocol for cache coherence.
  • Supports silent and debug simulation modes.
  • Provides detailed statistics, including cache hit/miss ratios and MESI state transitions.

Testing Strategies

  • Validates read and write operations.
  • Tests Pseudo-LRU functionality for efficient replacement.
  • Evaluates MESI FSM transitions for different cache scenarios.

Testing MESI protocol

image2

This project was undertaken as part of the ECE585 course at Portland State University under the guidance of Prof. Mark G. Faust.

About

A SystemVerilog-based simulation and design of a Last Level Cache (LLC) implementing the MESI protocol, featuring Pseudo-LRU replacement, multi-mode simulations, and comprehensive testing strategies for optimized processor performance.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •