의존성 설치 및 PyCUDA 설치
-
의존성 설치
sudo apt install python3-pip sudo pip3 install wget # 인터넷에서 파일 다운 받을 때 사용 sudo pip3 install Cython -
PyCUDA 설치
sudo apt install libboost-all-dev sudo apt install python3-numpy sudo apt install python3-matplotlib sudo apt install build-essential python-dev python-setuptools libboost-python-dev libboost-thread-dev -
pycuda 설치
~~~
cd Downloads
# 수동 설치 시
sudo wget https://files.pythonhosted.org/packages/5a/56/4682a5118a234d15aa1c8768a528aac4858c7b04d2674e18d586d3dfda04/pycuda-2021.1.tar.gz
cd pycuda-2021.1.1
sudo ./configure.py
make -j4
sudo python3 setup.py install
sudo pip3 install .
~~~
- LLVM 설치
~~~
wget https://github.com/llvm/llvm-project/releases/download/llvmorg-12.0.0/llvm-12.0.0.src.tar.xz
sudo tar -xvf llvm-12.0.0.src.tar.xz
cd llvm-12.0.0.src
sudo mkdir llvm_build_dir
cd llvm_build_dir/
cmake ../-DCMAKE_BUILD_TYPE=Release -DLLVM_TARGETS_TO_BUILD="ARM;X86;AArch64"
~~~
만약 CMake 3.x or higher is required. You are running version 3.x 식으로 cmake 의 최신버전이 아니여서 설치가 실패한다면
업데이트를 해주어야 한다.
~~~
sudo apt install libssl-dev # OpenSSL
wget https://github.com/Kitware/CMake/releases/download/v3.20.5/cmake-3.20.5.tar.gz
tar -xvf cmake-3.20.5.tar.gz
cd cmake-3.20.5
sudo ./bootstrap
make # 작동이 안된다면 sudo apt install make
sudo make install
~~~
설치가 잘 되었다면 cmake --version 명령어로 확인 할 수 있다.
이후 다시 cmake ../-DCMAKE_BUILD_TYPE=Release -DLLVM_TARGETS_TO_BUILD="ARM;X86;AArch64" 명령어를 입력해주자.
~~~
make -j4
sudo make install
cd bin/
echo "export LLVM_CONFIG=\""`pwd`"/llvm-config\"" >> ~/.bashrc
echo "alias llvm='"`pwd`"/llvm-lit'" >> ~/.bashrc source ~/.bashrc
sudo pip3 install llvmlite
~~~
