ECMAScript Native Parser Benchmark

Benchmark ECMAScript parsers implemented in native languages.

System

Property	Value
OS	Linux 6.14.0-1017-azure (x64)
CPU	AMD EPYC 7763 64-Core Processor
Cores	4
Memory	16 GB

Parsers

Yuku

Language: Zig

A high-performance & spec-compliant JavaScript/TypeScript compiler written in Zig.

Oxc

Language: Rust

A high-performance JavaScript and TypeScript parser written in Rust.

SWC

Language: Rust

An extensible Rust-based platform for compiling and bundling JavaScript and TypeScript.

Jam

Language: Zig

A JavaScript toolchain written in Zig featuring a parser, linter, formatter, printer, and vulnerability scanner.

Benchmarks

TypeScript

The TypeScript compiler source code bundled into a single file.

File size: 7.83 MB

Parser	Mean	Min	Max	Peak Memory (RSS)
Yuku	77.41 ms	76.82 ms	78.89 ms	39.7 MB
Oxc	77.74 ms	77.07 ms	79.55 ms	52.8 MB
SWC	171.39 ms	169.97 ms	173.07 ms	89.7 MB
Jam	201.19 ms	200.12 ms	203.77 ms	186.1 MB

Three.js

A popular 3D graphics library for the web.

File size: 1.96 MB

Parser	Mean	Min	Max	Peak Memory (RSS)
Oxc	17.23 ms	16.74 ms	17.87 ms	13.4 MB
Yuku	18.95 ms	18.59 ms	20.22 ms	10.3 MB
SWC	35.95 ms	34.59 ms	37.04 ms	22.4 MB
Jam	43.88 ms	43.40 ms	45.38 ms	39.5 MB

Ant Design

A popular React UI component library with enterprise-class design.

File size: 5.43 MB

Parser	Mean	Min	Max	Peak Memory (RSS)
Yuku	60.51 ms	59.70 ms	72.04 ms	30.5 MB
Oxc	61.86 ms	61.13 ms	62.74 ms	41.0 MB
SWC	130.79 ms	129.17 ms	132.30 ms	67.4 MB
Jam	Failed to parse	-	-	-

Run Benchmarks

Prerequisites

• Bun - JavaScript runtime and package manager
• Rust - For building Rust-based parsers
• Zig - For building Zig-based parsers (requires nightly/development version)
• Hyperfine - Command-line benchmarking tool

Steps

1. Clone the repository:

git clone https://github.com/yuku-toolchain/ecmascript-native-parser-benchmark.git
cd ecmascript-native-parser-benchmark

2. Install dependencies:

bun install

3. Run benchmarks:

bun bench

This will build all parsers and run benchmarks on all test files. Results are saved to the result/ directory.

Methodology

How Benchmarks Are Conducted

1. Build Phase: All parsers are compiled with release optimizations. Source files are embedded at compile time (Zig @embedFile, Rust include_str!) to eliminate file I/O from measurements:

   • Rust parsers: cargo build --release with LTO, single codegen unit, and symbol stripping
   • Zig parsers: zig build --release=fast

2. Benchmark Phase: Each parser is benchmarked using Hyperfine:

   • 100 warmup runs to ensure stable measurements
   • Multiple timed runs for statistical accuracy
   • Results exported to JSON for analysis

3. Measurement: Each benchmark measures the total time to:

   • Parse the entire file into an AST (source is embedded at compile time, no file I/O)
   • Clean up allocated memory

Test Files

The benchmark uses real-world JavaScript files from popular open-source projects to ensure results reflect practical performance characteristics.