crust

About

A small project to compare different features of Rust and C++ as a systems programming language.

Instructions to Compile

Rust installation

rustc is the compiler that is used. To install Rust on a Unix-like OS run the following command

curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

For all the ways to install rust please visit this site: https://www.rust-lang.org/tools/install

Code Compilation

Rust

All the Rust implementations are present in the directory src/rs/src. Run the following instructions to go to that directory from the root of this project and compile all the Rust codes

cd src/rs/src
./compile.sh

This will compile all the Rust src and put the executables in the directory src/rs/bin (path wrt project root)

cd ../bin

C++

All the C++ src files are present in the directory src/cxx. Run the following instructions to go to that directory from the root of this project and compile all the C++ codes

cd src/cxx
make

This will compile all the C++ src and put the executables in the directory src/cxx/bin (path wrt project root)

cd bin

Instructions to benchmark

The benchmark.py in the src directory, contains the Python script which runs the all the executables. It uses the config.json as a guide to run which executables. After it runs it stores the output in the file called results-{date}_{time}.json in the directory src/results (path wrt to project root).

config.json

Edit the config.json inside the src folder as follows.

To add an executable to compare with the others add an object like the following

{
    "binaries": [
    {...},
    {
        "name": "An easy name to identify in the result.json",
        "command": "Location of the binary along with its argument like how you would run from the terminal",
        "iterations": 10 //An integer value specifying the number of times the code will be run
    },
    {
        "name": "Linked list in Rust",
        "command": "rs/bin/linked_list",
        "iterations": 100
    }
    {...}
    ]
}

Running the benchmark

We used Python 3.5 version of Python. To run the benchmark simply run the following command

python benchmark.py

After executing all the binaries the result will be stored in the results-{date}_{time}.json file. It has the following structure

{
    "results": [
    {...},
    {
        "name": "The same you used in config.json",
        "iterations": 100, // An integer specifying the number of times the code was run
        "time_unit": "us", // Represents micro-seconds
        "mean": 1234.56, // mean time taken by the executable across all the runs
        "median": 1234.56, // the median of all the times taken by the executable
        "stddiv": 123.56 // the standard deviation of execution times across all the runs
    }
    {...}
    ]
}

Requirements

Tested on a new anaconda environment with Python 3.5, the only required package was psutil. Run the following command to install this requirement

pip install -r requirements.txt
# or simply
pip install psutil

If it shows any other dependency required just install them using pip

grakshith/crust