Get rid of warnings during the basic build step

BUDDY MLIR

An MLIR-based compiler framework designed for a co-design ecosystem from DSL (domain-specific languages) to DSA (domain-specific architectures). (Project page).

Getting Started

The default build system uses LLVM/MLIR as an external library. We also provide a one-step build strategy for users who only want to use our tools. Please make sure the dependencies are available on your machine.

LLVM/MLIR Dependencies

Before building, please make sure the dependencies are available on your machine.

Clone and Initialize

$ git clone [email protected]:buddy-compiler/buddy-mlir.git
$ cd buddy-mlir
$ git submodule update --init

Build and Test LLVM/MLIR/CLANG

$ cd buddy-mlir
$ mkdir llvm/build
$ cd llvm/build
$ cmake -G Ninja ../llvm \
    -DLLVM_ENABLE_PROJECTS="mlir;clang" \
    -DLLVM_TARGETS_TO_BUILD="host;RISCV" \
    -DLLVM_ENABLE_ASSERTIONS=ON \
    -DCMAKE_BUILD_TYPE=RELEASE
$ ninja check-mlir check-clang

If your target machine includes a Nvidia GPU, you can use the following configuration:

$ cmake -G Ninja ../llvm \
    -DLLVM_ENABLE_PROJECTS="mlir;clang" \
    -DLLVM_TARGETS_TO_BUILD="host;RISCV;NVPTX" \
    -DMLIR_ENABLE_CUDA_RUNNER=ON \
    -DLLVM_ENABLE_ASSERTIONS=ON \
    -DCMAKE_BUILD_TYPE=RELEASE

To enable MLIR Python bindings, please use the following configuration:

$ cmake -G Ninja ../llvm \
    -DLLVM_ENABLE_PROJECTS="mlir;clang" \
    -DLLVM_TARGETS_TO_BUILD="host;RISCV" \
    -DLLVM_ENABLE_ASSERTIONS=ON \
    -DCMAKE_BUILD_TYPE=RELEASE \
    -DMLIR_ENABLE_BINDINGS_PYTHON=ON \
    -DPython3_EXECUTABLE=$(which python3)

If your target machine has lld installed, you can use the following configuration:

$ cmake -G Ninja ../llvm \
    -DLLVM_ENABLE_PROJECTS="mlir;clang" \
    -DLLVM_TARGETS_TO_BUILD="host;RISCV" \
    -DLLVM_USE_LINKER=lld \
    -DLLVM_ENABLE_ASSERTIONS=ON \
    -DCMAKE_BUILD_TYPE=RELEASE

Build buddy-mlir

If you have previously built the llvm-project, you can replace the $PWD with the path to the directory where you have successfully built the llvm-project.

$ cd buddy-mlir
$ mkdir build
$ cd build
$ cmake -G Ninja .. \
    -DMLIR_DIR=$PWD/../llvm/build/lib/cmake/mlir \
    -DLLVM_DIR=$PWD/../llvm/build/lib/cmake/llvm \
    -DLLVM_ENABLE_ASSERTIONS=ON \
    -DCMAKE_BUILD_TYPE=RELEASE
$ ninja
$ ninja check-buddy

To utilize the Buddy Compiler Python package, please ensure that the MLIR Python bindings are enabled and use the following configuration:

$ cmake -G Ninja .. \
    -DMLIR_DIR=$PWD/../llvm/build/lib/cmake/mlir \
    -DLLVM_DIR=$PWD/../llvm/build/lib/cmake/llvm \
    -DLLVM_ENABLE_ASSERTIONS=ON \
    -DCMAKE_BUILD_TYPE=RELEASE \
    -DBUDDY_MLIR_ENABLE_PYTHON_PACKAGES=ON \
    -DPython3_EXECUTABLE=$(which python3)

If you want to add domain-specific framework support, please add the following cmake options:

Framework	Enable Option	Other Options
OpenCV	`-DBUDDY_ENABLE_OPENCV=ON`	Add `-DOpenCV_DIR=</PATH/TO/OPENCV/BUILD/>` or install OpenCV release version on your local device.

One-step building strategy

If you only want to use our tools and integrate them more easily into your projects, you can choose to use the one-step build strategy.

$ cmake -G Ninja -Bbuild \
    -DCMAKE_BUILD_TYPE=Release \
    -DLLVM_ENABLE_PROJECTS="mlir;clang" \
    -DLLVM_TARGETS_TO_BUILD="host;RISCV" \
    -DLLVM_EXTERNAL_PROJECTS="buddy-mlir" \
    -DLLVM_EXTERNAL_BUDDY_MLIR_SOURCE_DIR="$PWD" \
    -DLLVM_ENABLE_ASSERTIONS=ON \
    -DCMAKE_BUILD_TYPE=RELEASE \
    llvm/llvm
$ cd build
$ ninja check-mlir check-clang
$ ninja
$ ninja check-buddy

Use nix

This repository have nix flake support. You can follow the nix installation instruction and enable the flake features to have nix setup.

If you want to contribute to this project:

nix develop .

This will setup a bash shell with clang, clangd, cmake, ninja, and other necessary dependencies to build buddy-mlir from source.

If you want to use the buddy-mlir bintools

nix build .#buddy-mlir
./result/bin/buddy-opt --version

Dialects

Bud Dialect

Bud dialect is designed for testing and demonstrating.

DIP Dialect

DIP dialect is designed for digital image processing abstraction.

Tools

buddy-opt

The buddy-opt is the driver for dialects and optimization in buddy-mlir project.

buddy-lsp-server

This program should be a drop-in replacement for mlir-lsp-server, supporting new dialects defined in buddy-mlir. To use it, please directly modify mlir LSP server path in VSCode settings (or similar settings for other editors) to:

{
    "mlir.server_path": "YOUR_BUDDY_MLIR_BUILD/bin/buddy-lsp-server",
}

After modification, your editor should have correct completion and error prompts for new dialects such as rvv and gemmini.

Examples

The purpose of the examples is to give users a better understanding of how to use the passes and the interfaces in buddy-mlir. Currently, we provide three types of examples.

IR level conversion and transformation examples.
Domain-specific application level examples.
Testing and demonstrating examples.

For more details, please see the documentation of the examples.

How to Cite

If you find our project and research useful or refer to it in your own work, please cite our paper as follows:

@article{zhang2023compiler,
  title={Compiler Technologies in Deep Learning Co-Design: A Survey},
  author={Zhang, Hongbin and Xing, Mingjie and Wu, Yanjun and Zhao, Chen},
  journal={Intelligent Computing},
  year={2023},
  publisher={AAAS}
}

For direct access to the paper, please visit Compiler Technologies in Deep Learning Co-Design: A Survey.

buddy-compiler / buddy-mlir Goto Github PK

buddy-mlir's Introduction

BUDDY MLIR

Getting Started

LLVM/MLIR Dependencies

Clone and Initialize

Build and Test LLVM/MLIR/CLANG

Build buddy-mlir

One-step building strategy

Use nix

Dialects

Bud Dialect

DIP Dialect

Tools

buddy-opt

buddy-lsp-server

Examples

How to Cite

buddy-mlir's People

Contributors

Stargazers

Watchers

Forkers

buddy-mlir's Issues

Why it has to be i32 operands

Applying integer conversion in lowering process seems not to work

Personal suggestion

​when rum this commard:

Error:

Is your feature request related to a problem? Please describe.

Describe the solution you'd like

Recommend Projects

Recommend Topics

Recommend Org

when rum this commard: