Project overview

This repository contains a configurable cache and memory hierarchy simulator written in Python, along with MATLAB scripts used to generate publication-quality plots for an Advanced Computer Organization & Architecture assignment on cache performance analysis. The work explores how cache size, block size, associativity, and replacement policy affect hit ratio and average memory access time (AMAT) in single-level and multi-level cache designs.

Features

• Parameterized cache model supporting multiple replacement policies: LRU, FIFO, Random, and LFU.
• Support for sweeps over cache size, block size, and associativity, with per-cache and overall statistics such as hits, misses, hit ratio, miss ratio, AMAT, and total access time.
• Multi-level cache hierarchy modeling with configurable hit times and main memory access latency.
• CSV export of simulation results for downstream analysis and plotting.[2]
• MATLAB plotting script to generate AMAT comparison, hit ratio comparison, associativity effect, and replacement policy comparison figures using actual simulation outputs.

Repository structure

• assingment.py: Core cache & memory hierarchy simulator, including cache line and cache classes, hierarchy management, trace generators, and experiment helpers for running configurations and exporting CSVs.
• ACOA_plots.m: MATLAB script that loads simulation outputs and produces plots for AMAT vs configuration, L1 hit ratio, associativity effects, and replacement policy comparisons, saving them as PNG files.
• ACOA_FinalReport.pdf: Report documenting theoretical background, experiment design, simulation methodology, result analysis, and conclusions on cache performance optimization.

Requirements

• Python 3.x with standard library modules (math, random, time, csv, typing).
• MATLAB (R2013b or later recommended for readtable) for running the plotting script.

How to run simulations

1. Prepare or modify a memory access trace using the trace generator utilities or your own trace inside assingment.py.
2. Define one or more cache configurations (e.g., 1-level direct-mapped, 4-way set-associative, 2-level hierarchy) by instantiating Cache objects and passing them to run_simulation.
3. Run the Python script to execute the simulation, print per-level statistics, and write CSV files (e.g., for configuration sweeps and policy comparisons).
4. Ensure that the generated CSV files (such as policy_comparison.csv and associativity/configuration result files) are stored in the same directory as the MATLAB script if you intend to plot them.

How to generate plots

1. Run the Python simulations and confirm that the required CSV files with hit ratios and AMAT values have been generated.
2. Open ACOA_plots.m in MATLAB and adjust any file names or hard-coded values if your CSV naming differs.
3. Execute ACOA_plots.m to generate and save:
   • amat_comparison_plot_actual.png
   • hit_ratio_comparison_plot_actual.png
   • associativity_effect_plot_actual.png
   • replacement_policy_comparison.png

Academic context and usage

The simulator and plots were developed for a 2nd-year Advanced Computer Organization & Architecture course project titled “Performance Analysis and Design Strategies for Cache Memory Systems: A Simulation Approach.” The code is suitable as a learning tool or baseline for further research on cache hierarchy design, replacement policies, and performance tuning in modern processors.