A WebAssembly benchmarking suite, built on top of the Sightglass suite.
This suite is not a rigorous benchmark across WebAssembly engines; the intention is to provide a simple approximation and a rough comparison between compilation and execution times of various WebAssembly modules using different WebAssembly compilers (Baseline, Optimizing).
- Optimizing: Cranelift (Wasmtime)
- Baseline: Liftoff (V8)
- Baseline: RabaldrMonkey (SpiderMonkey)
These benchmarks exist, to answer the following:
- Approximate compilation time gains when using a baseline compiler
- Approximate execution time difference when using a baseline compiler
This suite doesn't use Sightglass's CLI to run the benchmarks. Instead it uses Criterion for Rust benchmarks and JavaScript's Performance
API for JavaScript benchmarks.
Sightglass uses Wasmtime's Benchmark API which intentionally uses a Shuffling Allocator to ensure statistically meaningful results, but make it unsuitable for comparisons between different engines.