A high performance WebAssembly interpreter written in C.

∼ 15.8x faster than other wasm interpreters (wac, life, wasm-micro-runtime)
∼ 4..5x slower than state of the art wasm JIT engines (liftoff, cranelift)
∼ 11.5x slower than native execution
_{* Based on CoreMark 1.0 benchmark. Your mileage may vary.}

Here's an online demo and a small getting started guide:

wasm3 passes the WebAssembly spec testsuite and is able to run many WASI apps.

Minimum useful system requirements: ~64Kb for code and ~10Kb RAM

wasm3 runs on a wide range of platforms:

• Linux, Windows, OS X
• Android, iOS
• OpenWRT-enabled routers
• Raspberry Pi, Orange Pi and other SBCs
• MCUs:
  Arduino MKR*, Arduino Due, Particle Photon,
  ESP8266, ESP32, Air602 (W600), nRF52, nRF51,
  Blue Pill (STM32F103C8T6), MXChip AZ3166 (EMW3166),
  Maix (K210), HiFive1 (E310), Fomu (ICE40UP5K), ATmega1284 etc.
• Browsers... yes, using WebAssembly itself!
• wasm3 can execute wasm3 (self-hosting)

wasm3 is built on top of Steven Massey's novel interpreter topology, with:

• Wasm 1.0 spec conformance
• WASI support
• Support of x86, x64, ARM, MIPS, RISC-V, Xtensa architectures

Why using "slow interpreter", instead of "fast JIT"?

In many situations, speed is not the main concern. Runtime executable size, code density, memory usage, startup latency can be all improved with the interpreter approach. Portability and security are much easier to achieve and maintain. Also on some platforms (i.e. iOS and WebAssembly itself) you can't generate executable code pages in runtime (so JIT is unavailable).

See DEV.md

See test/README.md

See PERFORMANCE.md

This project is released under The MIT License (MIT)