Concurrent Skiplist Project

Report can be found under https://mkleinegger.github.io/concurrent-skip-list/report.pdf

Overview

This project implements and benchmarks different skiplist variants for concurrent multiprocessor systems. The results of the benchmarks are visualized through plots and compiled into a report.

Directory Structure

.
├── Makefile               # Automates tasks like compiling, running benchmarks, and generating plots
├── README.md              # Documentation for the project
├── benchmark.py           # Main benchmarking script
├── benchmark_small.py     # Benchmark script for smaller test cases
├── benchmark_small_plots.py # Generates plots for small benchmarks
├── data/*                 # Contains benchmark results in organized directories
├── notebooks/             # Jupyter notebooks for visualizations
│   └── plots.ipynb        # Interactive analysis and plot generation
├── plots/*                # Generated plots from the benchmark scripts
├── report/                # Contains the LaTeX source for the generated report
│   └── report.tex         # Main LaTeX file
├── requirements.txt       # Python dependencies for the project
├── run_nebula.sh          # Script to run benchmarks on Nebula
├── run_setup_python.sh    # Script to set up a python environment
└── src/*                  # Source code for the skiplist and utilities

How to Run the Project

1. Run Benchmarks

To generate benchmark data, use the provided scripts included in the Makefile:

make bench-seq
make bench-global
make bench-fine
make bench-lockfree

Using those commands in combination with slurm on nebula should produce all results

2. Generate Report

If you want to generate a small sample using our small benchmark, you need to do the following:

1. Set $USERNAME in run_nebula.sh to your username on the nebula system
2. Run make zip
3. Run bash run_nebula.sh project.zip small-bench
4. Run make small-plot
5. Run make report

This should allow to generate the necessary report. If your python environment does not have all necessary requierements, we either provide those within ./requirements.txt to install into your environment or create and use a new environment with run_setup_python.sh. However for later approach you need to activate it first!

3. Generate Plots

To generate all Plots possibly found in the Report, simply use this created environment as a Jupyter kernel to run the notebook ./notebook/plots.ipynb, which generates all plots.

Additional Information

To reproduce the graphs we inkluded the gathered data from our runs under the ./data, such that it can be looked at and used for reproduction purposes.