Skip to content

Latest commit

 

History

History
114 lines (79 loc) · 2.52 KB

README.en.md

File metadata and controls

114 lines (79 loc) · 2.52 KB
Raw

WASM-AoT

Introduction

An AOT runtime that can run wasm in AVR architecture

Software architecture

app

Wasm sample application code Including benchmarks

bin

Generated executable files

include

Header files

libs

Third party Libraries Including Avrora simulator

runtime-benchs

Other runtime test sets

src

Core runtime code

utils

Data statistics tools

Installation tutorial

  1. Possible dependencies (Ubuntu system)
apt install git build-essential wget curl unzip python3 gcc-multilib xxd cmake wabt
apt install gcc-avr avr-libc avrdude gdb-avr openjdk-11-jdk

Instructions for use

  1. Compile wasm code [optional]

First write your test.wat file (writing wasm code directly in text format to facilitate debugging)

Then use the following command to convert it to .h file format(c header file)

wat2wasm test. wat -o test. wasm --enable-annotations -v
wasm2wat test. wasm -o testi. wat
xxd -i test. wasm > test. wasm. h

The additional parameters in the first command make it easy to view the details of wasm binary format

The second command reverses the conversion from binary format to text format, which can be used to verify the correctness of manually written code

To test whether the .wat file is correct, run python3 instance.py under app/wasmer-test directory (dependency: pip install wasmer wasmer_compiler_cranelift)

You can view the results directly.

If there is an import function in .wat file, please add it in instance.py.

  1. Compile code

Execute the following shell command in the root directory:

rm -rf build
mkdir build&&cd build
cmake ..
make
  1. Run in simulator

Execute the following shell command in the root directory:

cd libs
sh test.sh

Run benchmark

  1. Compile benchmark

Run in the app/benchmark folder

bash compile.sh

The script will automatically compile the benchmark (from C code to native elf and wasm files)

  1. Run benchmark

Run in the libs directory

bash wasmbench.sh

You can automatically run all wasm files generated by the previous compilation and output the results to [benchname].wasmoutput.txt file

  1. Run benchmark in native c

Run in the libs directory

bash nativebench.sh

You can automatically run all elf files generated in the first step of compilation and output the results to [benchname].nativeoutput.txt file, which is used to compare the running efficiency of runtime

Contribution

  1. Fork the repository
  2. Create Feat_xxx branch
  3. Commit your code
  4. Create Pull Request