Tryout Rust, Java on a CLI (no network)

1. Purpose of this blog post

When I published my last blog post, I got some feedback to improve the Rust version and I decided to remove the network factor.

Instead of calling an API to get some UUID, we will indicate a file and iterate through it to look for the first N items.

The parameters of the executable are :

PATH : path of the file containing the uuid
COUNT : number of uuid to request

If you want to try out the code, you can check the project on GITHUB.

The results (compile time, execution time) have been obtained on a MacBook Pro (2021) M1 Max.

I will do a Graal VM version to compare.

2. Contribution on uuid-rust (the sync version)

First I want to say, that I got a contribution from Francis Lalonde on the code from last post to add a sync version equivalent to the rust code. If you want to check the pull request.

And thanks to this contribution, I discovered a command-line benchmarking tool : hyperfine.

The code consists of a main file.

use std::str::FromStr;

fn main() -> Result<(), anyhow::Error> {
    let args: Vec<String> = std::env::args().collect();
    let version = u32::from_str(&args[1])?;
    let count = u32::from_str(&args[2])?;

    let url = format!("https://www.uuidtools.com/api/generate/v{}/count/{}", version, count);
    let uuids: Vec<String> = ureq::get(&url)
        .call()?
        .into_json()?;

    uuids.iter().for_each(|uuid| println!("{}", uuid));
    Ok(())
}

The sync Rust version is slightly better compared to the Java native executable.

Rust

uuid/uuid-rust-sync [🌱 master][📦 v0.1.0][ v1.68.0]
❯ hyperfine --runs 50 './target/release/uuid-rust-sync 4 5'

Benchmark 1: ./target/release/uuid-rust-sync 4 5
Time (mean ± σ):     132.7 ms ±  16.4 ms    [User: 4.0 ms, System: 4.4 ms]
Range (min … max):   109.4 ms … 218.9 ms    50 runs

Graal VM

❯ hyperfine --runs 50 './target/uuid-java 4 5'

Benchmark 1: ./target/uuid-java 4 5
  Time (mean ± σ):     136.8 ms ±  10.0 ms    [User: 14.1 ms, System: 9.5 ms]
  Range (min … max):   112.9 ms … 154.4 ms    50 runs

So let’s try to remove the network factor and open a file to look for a number of uuid in a file.

3. uuid-rust

The code consists of a main file. The important thing is that I use serde for deserialization.

use std::fs;
use std::str::FromStr;

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let args: Vec<String> = std::env::args().collect();
    let file_path = &args[1];
    let count = usize::from_str(&args[2])?;

    let content = fs::read_to_string(file_path);

    let uuids: Vec<String> = serde_json::from_str(&content?)?;

    uuids.iter().take(count).for_each(|uuid| println!("{}", uuid));

    Ok(())
}

4. uuid-java

The code consists of a main class. The important thing is that I use the Jackson library to deserialize my response into a list of UUID.

package com.xavierbouclet;

import com.fasterxml.jackson.core.type.TypeReference;
import com.fasterxml.jackson.databind.ObjectMapper;

import java.nio.file.Files;
import java.nio.file.Paths;
import java.util.Arrays;
import java.util.List;
import java.util.UUID;

public class App {

    public static void main(String[] args) throws Exception {
        var filepath = args[0];
        var count = Long.parseLong(args[1]);
        ObjectMapper objectMapper = new ObjectMapper();

        objectMapper.readValue(Files.readString(Paths.get(filepath)), new TypeReference<List<UUID>>() {
        }).stream().limit(count).forEach(
                System.out::println
        );
    }
}

5 Let’s run some tests

The tests consist of calling our executable with a json file and requesting a number of uuids in it.

Example of file :

[
  "40aef568-bca6-4b15-a367-f15964cab651",
  "6e3f166c-eff4-41b4-8b3d-c72116bf6f2d",
  "07fd7a9e-0e1e-45fc-a202-72ed73902b56",
  "e96128f9-1ec0-490f-85eb-9e6e9e07f47f",
  "5d5c9320-d610-42fc-8cf9-710158669749"
]

The file uuids.json contains 100 uuids and the files uuids2.json contains 83162 uuids.

Let’s run some tests.

Rust

hyperfine --runs 100  './uuid-rust/target/release/uuid-rust "/Users/xavierbouclet/Sources/uuid/uuids.json" 5'

Benchmark 1: ./uuid-rust/target/release/uuid-rust "/Users/xavierbouclet/Sources/uuid/uuids.json" 5
Time (mean ± σ):       0.6 ms ±   0.5 ms    [User: 0.5 ms, System: 0.3 ms]
Range (min … max):     0.1 ms …   2.6 ms    100 runs

5 items 83k file

hyperfine --runs 100  './uuid-rust/target/release/uuid-rust "/Users/xavierbouclet/Sources/uuid/uuids2.json" 5'

Benchmark 1: ./uuid-rust/target/release/uuid-rust "/Users/xavierbouclet/Sources/uuid/uuids2.json" 5
Time (mean ± σ):       7.9 ms ±   1.2 ms    [User: 6.0 ms, System: 1.7 ms]
Range (min … max):     6.8 ms …  13.9 ms    100 runs

1000 items 83k file

hyperfine --runs 100  './uuid-rust/target/release/uuid-rust "/Users/xavierbouclet/Sources/uuid/uuids2.json" 1000'
Benchmark 1: ./uuid-rust/target/release/uuid-rust "/Users/xavierbouclet/Sources/uuid/uuids2.json" 1000
Time (mean ± σ):       8.9 ms ±   0.4 ms    [User: 6.2 ms, System: 2.2 ms]
Range (min … max):     8.3 ms …  11.3 ms    100 runs

10000 items 83k file

hyperfine --runs 100  './uuid-rust/target/release/uuid-rust "/Users/xavierbouclet/Sources/uuid/uuids2.json" 10000'
Benchmark 1: ./uuid-rust/target/release/uuid-rust "/Users/xavierbouclet/Sources/uuid/uuids2.json" 10000
  Time (mean ± σ):      14.2 ms ±   0.7 ms    [User: 8.1 ms, System: 5.6 ms]
  Range (min … max):    13.5 ms …  18.8 ms    100 runs

40000 items 83k file

hyperfine --runs 100  './uuid-rust/target/release/uuid-rust "/Users/xavierbouclet/Sources/uuid/uuids2.json" 40000'
Benchmark 1: ./uuid-rust/target/release/uuid-rust "/Users/xavierbouclet/Sources/uuid/uuids2.json" 40000
Time (mean ± σ):      31.6 ms ±   0.6 ms    [User: 14.2 ms, System: 16.8 ms]
Range (min … max):    30.9 ms …  33.8 ms    100 runs

80000 items 83k file

hyperfine --runs 100  './uuid-rust/target/release/uuid-rust "/Users/xavierbouclet/Sources/uuid/uuids2.json" 80000'
Benchmark 1: ./uuid-rust/target/release/uuid-rust "/Users/xavierbouclet/Sources/uuid/uuids2.json" 80000
Time (mean ± σ):      54.7 ms ±   0.8 ms    [User: 22.3 ms, System: 31.7 ms]
Range (min … max):    53.0 ms …  59.1 ms    100 runs

Graal VM

hyperfine --runs 100  './uuid-java/target/uuid-java "/Users/xavierbouclet/Sources/uuid/uuids.json" 5'

Benchmark 1: ./uuid-java/target/uuid-java "/Users/xavierbouclet/Sources/uuid/uuids.json" 5
Time (mean ± σ):      10.3 ms ±   1.1 ms    [User: 4.0 ms, System: 3.5 ms]
Range (min … max):     9.0 ms …  13.7 ms    100 runs

5 items 83k file

hyperfine --runs 100  './uuid-java/target/uuid-java "/Users/xavierbouclet/Sources/uuid/uuids2.json" 5'

Benchmark 1: ./uuid-java/target/uuid-java "/Users/xavierbouclet/Sources/uuid/uuids2.json" 5
Time (mean ± σ):      39.1 ms ±   2.2 ms    [User: 29.7 ms, System: 6.0 ms]
Range (min … max):    36.6 ms …  45.7 ms    100 runs

1000 items 83k file

hyperfine --runs 100  './uuid-java/target/uuid-java "/Users/xavierbouclet/Sources/uuid/uuids2.json" 1000'
Benchmark 1: ./uuid-java/target/uuid-java "/Users/xavierbouclet/Sources/uuid/uuids2.json" 1000
  Time (mean ± σ):      40.0 ms ±   1.0 ms    [User: 30.2 ms, System: 6.1 ms]
  Range (min … max):    38.8 ms …  43.6 ms    100 runs

10000 items 83k file

hyperfine --runs 100  './uuid-java/target/uuid-java "/Users/xavierbouclet/Sources/uuid/uuids2.json" 10000'
Benchmark 1: ./uuid-java/target/uuid-java "/Users/xavierbouclet/Sources/uuid/uuids2.json" 10000
Time (mean ± σ):      49.9 ms ±   1.2 ms    [User: 36.2 ms, System: 9.8 ms]
Range (min … max):    48.5 ms …  53.2 ms    100 runs

40000 items 83k file

hyperfine --runs 100  './uuid-java/target/uuid-java "/Users/xavierbouclet/Sources/uuid/uuids2.json" 40000'
Benchmark 1: ./uuid-java/target/uuid-java "/Users/xavierbouclet/Sources/uuid/uuids2.json" 40000
Time (mean ± σ):      81.9 ms ±   1.4 ms    [User: 55.8 ms, System: 22.0 ms]
Range (min … max):    79.8 ms …  86.2 ms    100 runs

80000 items 83k file

hyperfine --runs 100  './uuid-java/target/uuid-java "/Users/xavierbouclet/Sources/uuid/uuids2.json" 80000'
Benchmark 1: ./uuid-java/target/uuid-java "/Users/xavierbouclet/Sources/uuid/uuids2.json" 80000
Time (mean ± σ):     123.2 ms ±   1.5 ms    [User: 81.7 ms, System: 37.6 ms]
Range (min … max):   121.3 ms … 131.7 ms    100 runs

We can see on the following graph that Rust is more performant than Java with Graal VM.

6. Conclusion

Rust overall is more performant than Java with Graal VM. And we can see, that more item we output it says consistant.

If you want to check the code on GitHub.

Follow Me

← Previous Post Next Post →