An experimental SIMD-accelerated HTML parser written in Go.
Built for fun, speed, and possibly breaking everything.
This is a fast, hand-crafted HTML parser that attempts to leverage SIMD (Single Instruction, Multiple Data) operations for blazing-fast parsing performance.
Heavily inspired by
simdjson— a ridiculously fast JSON parser that showed what's possible with SIMD. This project tries to apply similar ideas to HTML parsing, with varying degrees of success (and failure).
- I wanted to mess around with SIMD.
- I wanted to see how far I could push Go for this kind of task.
- I have a weird love for performance tuning.
- Parses some HTML using SIMD
- Written in Go, no C bindings
- Uses Plan9 assembly
- Uses bitwise operations
Pull requests are welcome, especially ones that fix my obvious mistakes. Open issues, file bugs, or just fork it and build your own better version.
This project is purely experimental and built for educational purposes. It is not production-ready and should not be relied upon in any critical system.
I take no responsibility for broken sites or corrupted HTML.
MIT – use it, break it, learn from it.