compression-sorts

A benchmark stand for testing permute algorithms in terms of optimal data compression

How To Install (Linux Ubuntu 22.04+)

First of all download repository:

git clone https://github.com/ElderlyPassionFruit/compression-sorts.git

After that initialize everything.

P.S. There and then, all commands must be executed strictly in the compression-sorts directory.

cd compression-sorts
./setup.sh

Yoo-hoo! Now you are ready to use compression-sorts!

Benchmarks

There are two benchmarks in project.

1. You can use my native implementation of all algorithms, just execute

   ./bin/benchmark
   

2. You can check how good current clickhouse implementation works (it must supports allow_experimental_optimized_row_order setting).

   If you want to do this, first of all, install and run clickhouse-server (for example you can use my PR, use this instruction).

   After that just execute

   python3 clickhouse_tests/main.py
   

How can you look at the test results?

On your own hardware

First, run the benchmarks from the previous paragraph. After that, tables with tests results (in .csv format) will be located in the tests_results directory.

Fill free to use research.ipynb jupyter notebook.

If you just execute both benchmarks and after that execute Run All, the program will draw many graphs showing the current efficiency of the algorithms on various tests.

On my hardware

You can also look at the results of tests performed on my hardware.

They are in research_with_results branch, in the notebook research.ipynb and in the tests_results directory.

Characteristics of my hardware:

• Processor: 8-core Intel(R) Xeon(R) Platinum 8259CL CPU @ 2.50GHz
• RAM: 30GB
• OS Version: Ubuntu 22.04.4 LTS