Skip to content
This repository has been archived by the owner on Apr 4, 2023. It is now read-only.

Latest commit

 

History

History
78 lines (55 loc) · 2.28 KB

README.md

File metadata and controls

78 lines (55 loc) · 2.28 KB

Snarkify WASM MSM Submission Rust Crate

Build

WASM package suitable for use with Webpack can be built with:

wasm-pack build --release

Resulting package will be available in the pkg directory including a Wasm module and JavaScript and TypeScript bindings. This package can be imported into Webpack applications.

Dependencies:

Benchmark

Wasm in Browser

./serve.sh

Opening https://localhost:8080 in your browser will give a page where you can provide input test vector files to run. You can generate input files with:

cargo run --bin generate-input-files

Native

cargo bench

Note: Default features in this crate are intended for Wasm targets. If looking to run optimally on native, turn off the default features for compilation.

Benchmarking notes

  • On x86 Linux, more stable benchmarking results can be achieved by increasing the execution pririty of the benchmark via nice. Here is an example command to do that for native CPU profiling.

    nice -n -15 cargo bench

  • It's possible to get pprof profile outputs by running the benchmarks with the following command.

    # Run 30 seconds of profiling for each benchmark target. Use nice to increase priority.
nice -n -15 cargo bench --bench bench_pippenger_msm -- --profile-time 30

    Profile outputs will be stored at ./target/criterion/msm/$FUNCTION/$INPUT_SIZE/profile/profile.pprof Profiles can be opened with the pprof tool, which can be installed and run with

    go tool pprof -http localhost:8080 target/criterion/msm/$FUNCTION/$INPUT_SIZE/profile.pprof

  • Especially when trying to meausre smaller effects, 100 samples may not be enough. Number of samples can be increased using the --sample-size flag.

    nice -n -15 cargo bench --bench bench_pippenger_msm -- --sample-size 1000

Note: SUPERCOP, which is a cryptography benchmarking suite, has useful recommendations on reducing randomness in benchmarking results, mostly focused on randomness due to clock rate variability.

https://bench.cr.yp.to/supercop.html