mst

A demo of Kruskal's algorithm for Minimum Spanning Tree in JS and Rust(wasm).

https://mst.boopathi.in/

The graph

Nodes: Generate n random points in the full screen canvas
Edges: Every node is connected to every other node. n * (n-1) / 2 edges
Weights: The weight of an edge is the distance between the two nodes it connects

Try it locally

Requirements

Rust > 1.30.0
- curl https://sh.rustup.rs -sSf | sh
wasm-pack
- cargo install wasm-pack
configure wasm32 target for rust
- rustup target add wasm32-unknown-unknown

Build

The rust library mst needs to be built using wasm-pack with the config to output for web (this target takes care of fetching of the wasm file and the WebAssembly instantiation).

wasm-pack build --target web

Run

After the wasm files are built you can verify it by looking into the pkg/ directory, the following files -

pkg/mst.js
pkg/mst_bg.wasm

which will be fetched from main.js

Simply start a static file server to serve index.html

python -m SimpleHTTPServer

Findings

Rust performs ~3x-4x faster for bigger graphs. For smaller ones, the performance improvement is not worth the effort - numNodes=100 gives 8ms(JS) to 6ms(Rust) improvement.

[RUST] Took ~137ms to send input, construct Graph in Rust and compute MST for 1000 nodes with 499500 edges and serialize in Rust and deserialize in JS 999 edges
[JS] Took ~451ms to compute MST for 1000 nodes with 499500 edges

License

Licensed under either of

Apache License, Version 2.0, (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)

at your option.

Contribution

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.

boopathi / mst

mst