nliver / blue-porcelain

Blue-porcelain GPGPU

Blue-porcelain is a GPGPU design proposed by the Advanced Architecture Laboratory of Shanghai Jiao Tong University.

Directory structure

• driver: Host drivers repository.
• runtime: Kernel Runtime software.
• sim: Simulators repository.
• benchmarks: Benchmarks for testing.
• ci: Continuous integration scripts.

Build

Supported OS Platforms

• Ubuntu 18.04

Toolchain Dependencies

• gcc 7.5.0
• g++ 7.5.0
• GNU Make 4.1
• CUDA 10.2 Runtime

Install development tools

sudo apt-get install build-essential
sudo apt-get install git bc numdiff

Install codebase

git clone https://github.com/SJTU-ACALab/blue-porcelain.git

Setup Environment

Make sure that the CUDA_INSTALL_PATH is set to the location (e.g., /usr/local/cuda) where the CUDA Toolkit is installed and the $CUDA_INSTALL_PATH/bin is in your PATH. You can add the following instructions to the .bashrc file (assume the CUDA Toolkit is installed in /usr/local/cuda):

export CUDA_INSTALL_PATH=/usr/local/cuda
export PATH=$CUDA_INSTALL_PATH/bin:$PATH

Build sources

source setup_environment
make -s

Run app

Before running a simulation for an application, make sure you have set up the environment. Then, just simply input the application execution command.

source setup_environment
./application

Run ci benchmark test

You can run our benchmark automatically under the ci directory. Make sure you have set up the environment. Also, the integration test is supported by Python. Please install Python(3.6 ~ 3.9 is available) if you want to use this tool.

$ source setup_environment
$ cd ./ci
$ sh run.sh -a

-core: set the sm core number(default is 1)

-a : build both the simulation program and benchmark program

-b : only build the simulation program

-bb : only build the benchmark program

Docker

Build a docker image and run the regressions in the docker.

cd blue-porcelain
docker build -t acalab/gpgpu .
docker run -w /root/blue-porcelain/ci -it --rm acalab/gpgpu bash run.sh -a

OpenGPU GPGPU Platform

For more information about our Open GPGPU Platform, go here: https://gpgpuarch.org

License

License information can be found in the LICENSE file.

Third party license information can be found in the THIRDPARTY file.