add auto install scripts

This commit is contained in:
xulei 2022-02-15 19:50:32 +08:00
parent 33c3a9e910
commit 38a1ed4f73
8 changed files with 633 additions and 260 deletions

View File

@ -1,67 +0,0 @@
#!/bin/bash
set -ex
MINDSPORE_VERSION=${MINDSPORE_VERSION:-1.5.0}
PYTHON_VERSION=${PYTHON_VERSION:-3.7.5}
#use huaweicloud mirror in China
sudo sed -i "s@http://.*archive.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
sudo sed -i "s@http://.*security.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
sudo apt-get update
# install python 3.7 and make it default
sudo apt-get install gcc-7 libgmp-dev curl python3.7 -y
sudo update-alternatives --install /usr/bin/python python /usr/bin/python3.7 100
cd /tmp
curl -O https://mirrors.tuna.tsinghua.edu.cn/anaconda/miniconda/Miniconda3-py37_4.10.3-Linux-x86_64.sh
bash Miniconda3-py37_4.10.3-Linux-x86_64.sh -b
# add conda to PATH
echo -e 'export PATH=~/miniconda3/bin/:$PATH' >> ~/.bash_profile
echo -e '. ~/miniconda3/etc/profile.d/conda.sh' >> ~/.bash_profile
source ~/.bash_profile
conda init bash
# setting up conda mirror with qinghua source
cat >~/.condarc <<END
channels:
- defaults
show_channel_urls: true
default_channels:
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/r
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/msys2
custom_channels:
conda-forge: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
msys2: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
bioconda: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
menpo: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
pytorch: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
pytorch-lts: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
simpleitk: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
END
#initialize conda env and install mindspore-cpu
conda create -n ms_${PYTHON_VERSION} python=${PYTHON_VERSION} -y
conda activate ms_${PYTHON_VERSION}
conda install mindspore-cpu=${MINDSPORE_VERSION} -c mindspore -c conda-forge -y
# check if it is the right mindspore version
python -c "import mindspore;mindspore.run_check()"
# check if it can be run with GPU
cat > example.py <<END
import numpy as np
from mindspore import Tensor
import mindspore.ops as ops
import mindspore.context as context
context.set_context(device_target="GPU")
x = Tensor(np.ones([1,3,3,4]).astype(np.float32))
y = Tensor(np.ones([1,3,3,4]).astype(np.float32))
print(ops.add(x, y))
END
python example.py

View File

@ -0,0 +1,94 @@
#!/bin/bash
# Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
# Prepare and Install mindspore cpu by conda on Ubuntu 18.04.
#
# This file will:
# - change deb source to huaweicloud mirror
# - install mindspore dependencies via apt like gcc, libgmp
# - install conda and set up environment for mindspore
# - install mindspore-cpu by conda
#
# Augments:
# - PYTHON_VERSION: python version to set up. [3.7(default), 3.9]
# - MINDSPORE_VERSION: mindspore version to install, default 1.6.0
#
# Usage:
# Run script like `bash ./ubuntu-cpu-conda.sh`.
# To set augments, run it as `PYTHON_VERSION=3.9 MINDSPORE_VERSION=1.5.0 bash ./ubuntu-cpu-conda.sh`.
set -e
PYTHON_VERSION=${PYTHON_VERSION:-3.7}
MINDSPORE_VERSION=${MINDSPORE_VERSION:-1.6.0}
available_py_version=(3.7 3.9)
if [[ " ${available_py_version[*]} " != *" $PYTHON_VERSION "* ]]; then
echo "PYTHON_VERSION is '$PYTHON_VERSION', but available versions are [${available_py_version[*]}]."
exit 1
fi
install_conda() {
conda_file_name="Miniconda3-py3${PYTHON_VERSION##*.}_4.10.3-Linux-$(arch).sh"
cd /tmp
wget https://mirrors.tuna.tsinghua.edu.cn/anaconda/miniconda/$conda_file_name
bash $conda_file_name -b
cd -
. ~/miniconda3/etc/profile.d/conda.sh
conda init bash
# setting up conda mirror with tsinghua source
cat >~/.condarc <<END
channels:
- defaults
show_channel_urls: true
default_channels:
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/r
- https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/msys2
custom_channels:
conda-forge: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
msys2: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
bioconda: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
menpo: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
simpleitk: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
END
}
# use huaweicloud mirror in China
sudo sed -i "s@http://.*archive.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
sudo sed -i "s@http://.*security.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
sudo apt-get update
sudo apt-get install gcc-7 libgmp-dev -y
# install conda
set +e && type conda &>/dev/null
if [[ $? -eq 0 ]]; then
echo "conda has been installed, skip."
source "$(conda info --base)"/etc/profile.d/conda.sh
else
install_conda
fi
set -e
# set up conda env and install mindspore-cpu
env_name=mindspore_py3${PYTHON_VERSION##*.}
conda create -n $env_name python=${PYTHON_VERSION} -y
conda activate $env_name
conda install mindspore-cpu=${MINDSPORE_VERSION} -c mindspore -c conda-forge -y
# check mindspore installation
python -c "import mindspore;mindspore.run_check()"

View File

@ -0,0 +1,69 @@
#!/bin/bash
# Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
# Prepare and Install mindspore cpu by pip on Ubuntu 18.04.
#
# This file will:
# - change deb source to huaweicloud mirror
# - install mindspore dependencies via apt like gcc, libgmp
# - install python3 & pip3 via apt and set it to default
# - install mindspore-cpu within new installed python by pip
#
# Augments:
# - PYTHON_VERSION: python version to install. [3.7(default), 3.9]
# - MINDSPORE_VERSION: mindspore version to install, default 1.6.0
#
# Usage:
# Need root permission to run, like `sudo bash ./ubuntu-cpu-pip.sh`.
# To set augments, run it as `sudo PYTHON_VERSION=3.9 MINDSPORE_VERSION=1.5.0 bash ./ubuntu-cpu-pip.sh`.
set -e
PYTHON_VERSION=${PYTHON_VERSION:-3.7}
MINDSPORE_VERSION=${MINDSPORE_VERSION:-1.6.0}
available_py_version=(3.7 3.9)
if [[ " ${available_py_version[*]} " != *" $PYTHON_VERSION "* ]]; then
echo "PYTHON_VERSION is '$PYTHON_VERSION', but available versions are [${available_py_version[*]}]."
exit 1
fi
declare -A version_map=()
version_map["3.7"]="${MINDSPORE_VERSION}-cp37-cp37m"
version_map["3.9"]="${MINDSPORE_VERSION}-cp39-cp39"
# use huaweicloud mirror in China
sed -i "s@http://.*archive.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
sed -i "s@http://.*security.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
apt-get update
apt-get install gcc-7 libgmp-dev -y
# python
add-apt-repository -y ppa:deadsnakes/ppa
apt-get install python$PYTHON_VERSION python$PYTHON_VERSION-distutils python3-pip -y
update-alternatives --install /usr/bin/python python /usr/bin/python$PYTHON_VERSION 100
# pip
sudo -u $SUDO_USER python -m pip install pip -i https://pypi.tuna.tsinghua.edu.cn/simple
update-alternatives --install /usr/bin/pip pip ~/.local/bin/pip$PYTHON_VERSION 100
update-alternatives --install /usr/local/bin/pip pip ~/.local/bin/pip$PYTHON_VERSION 100
sudo -u $SUDO_USER pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple
# install mindspore whl
arch=`uname -m`
sudo -u $SUDO_USER pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/${MINDSPORE_VERSION}/MindSpore/cpu/${arch}/mindspore-${version_map["$PYTHON_VERSION"]}-linux_${arch}.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple
# check mindspore installation
sudo -u $SUDO_USER python -c "import mindspore;mindspore.run_check()"

View File

@ -0,0 +1,81 @@
#!/bin/bash
# Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
# Prepare environment for mindspore cpu compilation on Ubuntu 18.04.
#
# This file will:
# - change deb source to huaweicloud mirror
# - install compile dependencies via apt like cmake, gcc
# - install python3 & pip3 via apt and set it to default
# - install LLVM if LLVM is set to on.
#
# Augments:
# - PYTHON_VERSION: python version to install. [3.7(default), 3.9]
# - LLVM: whether to install optional dependency LLVM for graph kernel fusion. [on, off(default)]
#
# Usage:
# Need root permission to run, like `sudo bash ./ubuntu-cpu-source.sh`.
# To set augments, run it as `sudo PYTHON_VERSION=3.9 LLVM=on bash ./ubuntu-cpu-source.sh`.
set -e
PYTHON_VERSION=${PYTHON_VERSION:-3.7}
LLVM=${LLVM:-off}
available_py_version=(3.7 3.9)
if [[ " ${available_py_version[*]} " != *" $PYTHON_VERSION "* ]]; then
echo "PYTHON_VERSION is '$PYTHON_VERSION', but available versions are [${available_py_version[*]}]."
exit 1
fi
# use huaweicloud mirror in China
sed -i "s@http://.*archive.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
sed -i "s@http://.*security.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
# base packages
apt-get update
apt-get install software-properties-common lsb-release -y
apt-get install curl tcl gcc-7 git libgmp-dev patch libnuma-dev -y
# cmake
wget -O - https://apt.kitware.com/keys/kitware-archive-latest.asc 2>/dev/null | apt-key add -
apt-add-repository "deb https://apt.kitware.com/ubuntu/ $(lsb_release -cs) main"
apt-get install cmake -y
# optional dependency LLVM for graph-computation fusion
if [[ X"$LLVM" == "Xon" ]]; then
wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key | apt-key add -
add-apt-repository "deb http://apt.llvm.org/bionic/ llvm-toolchain-bionic-12 main"
apt-get update
apt-get install llvm-12-dev -y
fi
# python
add-apt-repository -y ppa:deadsnakes/ppa
apt-get install python$PYTHON_VERSION python$PYTHON_VERSION-dev python$PYTHON_VERSION-distutils python3-pip -y
update-alternatives --install /usr/bin/python python /usr/bin/python$PYTHON_VERSION 100
# pip
sudo -u $SUDO_USER python -m pip install pip -i https://pypi.tuna.tsinghua.edu.cn/simple
update-alternatives --install /usr/bin/pip pip ~/.local/bin/pip$PYTHON_VERSION 100
update-alternatives --install /usr/local/bin/pip pip ~/.local/bin/pip$PYTHON_VERSION 100
sudo -u $SUDO_USER pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple
# wheel
sudo -u $SUDO_USER pip install wheel
# python 3.9 needs setuptools>44.0
sudo -u $SUDO_USER pip install -U setuptools
echo "The environment is ready to clone and compile mindspore."

View File

@ -1,26 +1,74 @@
#!/bin/bash #!/bin/bash
set -ex # Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
MINDSPORE_VERSION=${MINDSPORE_VERSION:-1.5.0} # Prepare and Install mindspore gpu by conda on Ubuntu 18.04.
PYTHON_VERSION=${PYTHON_VERSION:-3.7.5} #
MINDSPORE_VERSION=${MINDSPORE_VERSION:-1.5.0} # This file will:
CUDA_VERSION=${CUDA_VERSION:-11.1.1-1} # - change deb source to huaweicloud mirror
LIB_CUDA_VERSION=${LIB_CUDA_VERSION:-8.0.5.39-1+cuda11.1} # - install mindspore dependencies via apt like gcc, libgmp
DISTRIBUTED=${DISTRIBUTED:-false} # - install conda and set up environment for mindspore
CUDATOOLKIT_VERSION=${CUDATOOLKIT_VERSION:-11.1} # - install mindspore-gpu by conda
CUDNN_VERSION=${CUDNN_VERSION:-8.0.5} # - compile and install Open MPI if OPENMPI is set to on.
#
# Augments:
# - PYTHON_VERSION: python version to install. [3.7(default), 3.9]
# - MINDSPORE_VERSION: mindspore version to install, default 1.6.0
# - CUDA_VERSION: CUDA version to install. [10.1(default), 11.1]
# - OPENMPI: whether to install optional package Open MPI for distributed training. [on, off(default)]
#
# Usage:
# Run script like `bash -i ./ubuntu-gpu-conda.sh`.
# To set augments, run it as `PYTHON_VERSION=3.9 CUDA_VERSION=11.1 OPENMPI=on bash -i ./ubuntu-gpu-conda.sh`.
cd /tmp set -e
curl -O https://mirrors.tuna.tsinghua.edu.cn/anaconda/miniconda/Miniconda3-py37_4.10.3-Linux-x86_64.sh
bash Miniconda3-py37_4.10.3-Linux-x86_64.sh
# add conda to PATH PYTHON_VERSION=${PYTHON_VERSION:-3.7}
echo -e 'export PATH=~/miniconda3/bin/:$PATH' >> ~/.bash_profile MINDSPORE_VERSION=${MINDSPORE_VERSION:-1.6.0}
echo -e '. ~/miniconda3/etc/profile.d/conda.sh' >> ~/.bash_profile CUDA_VERSION=${CUDA_VERSION:-10.1}
source ~/.bash_profile OPENMPI=${OPENMPI:-off}
conda init bash
# setting up conda mirror available_py_version=(3.7 3.9)
cat >~/.condarc <<END if [[ " ${available_py_version[*]} " != *" $PYTHON_VERSION "* ]]; then
echo "PYTHON_VERSION is '$PYTHON_VERSION', but available versions are [${available_py_version[*]}]."
exit 1
fi
available_cuda_version=(10.1 11.1)
if [[ " ${available_cuda_version[*]} " != *" $CUDA_VERSION "* ]]; then
echo "CUDA_VERSION is '$CUDA_VERSION', but available versions are [${available_cuda_version[*]}]."
exit 1
fi
# add value to environment variable if value is not in it
add_env() {
local name=$1
if [[ ":${!name}:" != *":$2:"* ]]; then
echo -e "export $1=$2:\$$1" >> ~/.bashrc
fi
}
install_conda() {
conda_file_name="Miniconda3-py3${PYTHON_VERSION##*.}_4.10.3-Linux-$(arch).sh"
cd /tmp
wget https://mirrors.tuna.tsinghua.edu.cn/anaconda/miniconda/$conda_file_name
bash $conda_file_name -b
cd -
. ~/miniconda3/etc/profile.d/conda.sh
conda init bash
# setting up conda mirror with tsinghua source
cat >~/.condarc <<END
channels: channels:
- defaults - defaults
show_channel_urls: true show_channel_urls: true
@ -33,36 +81,57 @@ custom_channels:
msys2: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud msys2: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
bioconda: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud bioconda: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
menpo: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud menpo: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
pytorch: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
pytorch-lts: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
simpleitk: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud simpleitk: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
END END
}
#initialize conda env and install mindspore-cpu # use huaweicloud mirror in China
sudo sed -i "s@http://.*archive.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
sudo sed -i "s@http://.*security.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
sudo apt-get update
conda create -n ms_${PYTHON_VERSION} python=${PYTHON_VERSION} -y sudo apt-get install curl gcc-7 libgmp-dev -y
conda activate ms_${PYTHON_VERSION}
# install gmp 6.1.2, downloading gmp is slow # optional openmpi for distributed training
# echo "install gmp start" if [[ X"$OPENMPI" == "Xon" ]]; then
# sudo apt-get install m4 -y cd /tmp
# cd /tmp curl -O https://download.open-mpi.org/release/open-mpi/v4.0/openmpi-4.0.3.tar.gz
# curl -O https://gmplib.org/download/gmp/gmp-6.1.2.tar.xz tar xzf openmpi-4.0.3.tar.gz
# xz -d gmp-6.1.2.tar.xz cd openmpi-4.0.3
# tar xvzf gmp-6.1.2.tar && cd gmp-6.1.2 ./configure --prefix=/usr/local/openmpi-4.0.3
# ./configure --prefix=/usr/local/gmp-6.1.2 make
# make sudo make install
# sudo make install set +e && source ~/.bashrc
# echo 'export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/gmp-6.1.2/lib' >> ~/.bash_profile set -e
add_env PATH /usr/local/openmpi-4.0.3/bin
add_env LD_LIBRARY_PATH /usr/local/openmpi-4.0.3/lib
fi
# install mindspore-gpu with conda # install conda
conda install mindspore-gpu=${MINDSPORE_VERSION} cudatoolkit=${CUDATOOLKIT_VERSION} -c mindspore -c conda-forge -y set +e && type conda &>/dev/null
if [[ $? -eq 0 ]]; then
echo "conda has been installed, skip."
source "$(conda info --base)"/etc/profile.d/conda.sh
else
install_conda
fi
set -e
# check if it is the right mindspore version # set up conda env and install mindspore-cpu
env_name=mindspore_py3${PYTHON_VERSION##*.}
declare -A cudnn_version_map=()
cudnn_version_map["10.1"]="7.6.5"
cudnn_version_map["11.1"]="8.1.0"
conda create -n $env_name python=${PYTHON_VERSION} -y
conda activate $env_name
conda install mindspore-gpu=${MINDSPORE_VERSION} \
cudatoolkit=${CUDA_VERSION} cudnn=${cudnn_version_map[$CUDA_VERSION]} -c mindspore -c conda-forge -y
# check mindspore installation
python -c "import mindspore;mindspore.run_check()" python -c "import mindspore;mindspore.run_check()"
# check if it can be run with GPU # check if it can be run with GPU
cd /tmp
cat > example.py <<END cat > example.py <<END
import numpy as np import numpy as np
from mindspore import Tensor from mindspore import Tensor
@ -74,5 +143,5 @@ x = Tensor(np.ones([1,3,3,4]).astype(np.float32))
y = Tensor(np.ones([1,3,3,4]).astype(np.float32)) y = Tensor(np.ones([1,3,3,4]).astype(np.float32))
print(ops.add(x, y)) print(ops.add(x, y))
END END
python example.py
python example.py cd -

View File

@ -1,124 +1,144 @@
#!/bin/bash #!/bin/bash
set -ex # Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
PYTHON_VERSION=${PYTHON_VERSION:-3.7.5} # Prepare and Install mindspore gpu by pip on Ubuntu 18.04.
MINDSPORE_VERSION=${MINDSPORE_VERSION:-1.5.0} #
CUDA_VERSION=${CUDA_VERSION:-8.0.5.39-1+cuda11.1} # This file will:
DISTRIBUTED=${DISTRIBUTED:-false} # - change deb source to huaweicloud mirror
CUDA_VERSION=${CUDA_VERSION:-11.1.1-1} # - install mindspore dependencies via apt like gcc, libgmp
CUDA_INSTALL_PATH=${CUDA_INSTALL_PATH:-cuda-11.1} # - install python3 & pip3 via apt and set it to default
LIBNCCL2_VERSION=${LIBNCCL2_VERSION:-2.7.8-1+cuda11.1} # - install CUDA by run file and cudnn via apt.
ARCH=$(uname -m) # - install mindspore-cpu within new installed python by pip
# - compile and install Open MPI if OPENMPI is set to on.
#
# Augments:
# - PYTHON_VERSION: python version to install. [3.7(default), 3.9]
# - MINDSPORE_VERSION: mindspore version to install, default 1.6.0
# - CUDA_VERSION: CUDA version to install. [10.1(default), 11.1]
# - OPENMPI: whether to install optional package Open MPI for distributed training. [on, off(default)]
#
# Usage:
# Need root permission to run, like `sudo bash -i ./ubuntu-gpu-pip.sh`.
# To set augments, run it as `sudo PYTHON_VERSION=3.9 CUDA_VERSION=11.1 OPENMPI=on bash -i ./ubuntu-gpu-pip.sh`.
set -e
PYTHON_VERSION=${PYTHON_VERSION:-3.7}
MINDSPORE_VERSION=${MINDSPORE_VERSION:-1.6.0}
CUDA_VERSION=${CUDA_VERSION:-10.1}
OPENMPI=${OPENMPI:-off}
available_py_version=(3.7 3.9)
if [[ " ${available_py_version[*]} " != *" $PYTHON_VERSION "* ]]; then
echo "PYTHON_VERSION is '$PYTHON_VERSION', but available versions are [${available_py_version[*]}]."
exit 1
fi
available_cuda_version=(10.1 11.1)
if [[ " ${available_cuda_version[*]} " != *" $CUDA_VERSION "* ]]; then
echo "CUDA_VERSION is '$CUDA_VERSION', but available versions are [${available_cuda_version[*]}]."
exit 1
fi
declare -A minimum_driver_version_map=()
minimum_driver_version_map["10.1"]="418.39"
minimum_driver_version_map["11.1"]="450.80.02"
driver_version=$(modinfo nvidia | grep ^version | awk '{printf $2}')
if [[ $driver_version < ${minimum_driver_version_map[$CUDA_VERSION]} ]]; then
echo "CUDA $CUDA_VERSION minimum required driver version is ${minimum_driver_version_map[$CUDA_VERSION]}, \
but current nvidia driver version is $driver_version, please upgrade your driver manually."
exit 1
fi
cuda_name="cuda-$CUDA_VERSION"
declare -A version_map=() declare -A version_map=()
version_map["3.7.5"]="${MINDSPORE_VERSION}-cp37-cp37m" version_map["3.7"]="${MINDSPORE_VERSION}-cp37-cp37m"
version_map["3.9.0"]="${MINDSPORE_VERSION}-cp39-cp39m" version_map["3.9"]="${MINDSPORE_VERSION}-cp39-cp39"
#use huaweicloud mirror in China # add value to environment variable if value is not in it
sudo sed -i "s@http://.*archive.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list add_env() {
sudo sed -i "s@http://.*security.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list local name=$1
sudo apt-get update if [[ ":${!name}:" != *":$2:"* ]]; then
echo -e "export $1=$2:\$$1" >> ~/.bashrc
fi
}
# install python 3.7 and make it default # use huaweicloud mirror in China
sudo apt-get install gcc-7 libgmp-dev curl python3.7 openssl ubuntu-drivers-common openssl software-properties-common -y sed -i "s@http://.*archive.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
sudo update-alternatives --install /usr/bin/python python /usr/bin/python3.7 100 sed -i "s@http://.*security.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
apt-get update
apt-get install curl make gcc-7 libgmp-dev linux-headers-"$(uname -r)" -y
# python
add-apt-repository -y ppa:deadsnakes/ppa
apt-get install python$PYTHON_VERSION python$PYTHON_VERSION-distutils python3-pip -y
update-alternatives --install /usr/bin/python python /usr/bin/python$PYTHON_VERSION 100
# pip
sudo -u $SUDO_USER python -m pip install pip -i https://pypi.tuna.tsinghua.edu.cn/simple
update-alternatives --install /usr/bin/pip pip ~/.local/bin/pip$PYTHON_VERSION 100
update-alternatives --install /usr/local/bin/pip pip ~/.local/bin/pip$PYTHON_VERSION 100
sudo -u $SUDO_USER pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple
# install cuda/cudnn
cd /tmp cd /tmp
curl -O https://bootstrap.pypa.io/get-pip.py declare -A cuda_url_map=()
sudo python get-pip.py cuda_url_map["10.1"]=https://developer.download.nvidia.com/compute/cuda/10.1/Prod/local_installers/cuda_10.1.243_418.87.00_linux.run
cuda_url_map["11.1"]=https://developer.download.nvidia.com/compute/cuda/11.1.1/local_installers/cuda_11.1.1_455.32.00_linux.run
cuda_url=${cuda_url_map[$CUDA_VERSION]}
wget $cuda_url
sh ${cuda_url##*/} --silent --toolkit
cd -
apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
add-apt-repository "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/ /"
add-apt-repository "deb https://developer.download.nvidia.cn/compute/machine-learning/repos/ubuntu1804/x86_64/ /"
apt-get update
declare -A cudnn_name_map=()
cudnn_name_map["10.1"]="libcudnn7=7.6.5.32-1+cuda10.1 libcudnn7-dev=7.6.5.32-1+cuda10.1"
cudnn_name_map["11.1"]="libcudnn8=8.0.4.30-1+cuda11.1 libcudnn8-dev=8.0.4.30-1+cuda11.1"
apt-get install --no-install-recommends ${cudnn_name_map[$CUDA_VERSION]} -y
# add pip mirror # add cuda to path
mkdir -p ~/.pip set +e && source ~/.bashrc
cat > ~/.pip/pip.conf <<END set -e
[global] add_env PATH /usr/local/cuda/bin
index-url = https://repo.huaweicloud.com/repository/pypi/simple add_env LD_LIBRARY_PATH /usr/local/cuda/lib64
trusted-host = repo.huaweicloud.com add_env LD_LIBRARY_PATH /usr/lib/x86_64-linux-gnu
timeout = 120 set +e && source ~/.bashrc
END set -e
# install nvidia driver if not presented # optional openmpi for distributed training
# root@ecs-gpu-testing:~# ubuntu-drivers devices if [[ X"$OPENMPI" == "Xon" ]]; then
# == /sys/devices/pci0000:20/0000:20:00.0/0000:21:01.0 == cd /tmp
# modalias : pci:v000010DEd00001EB8sv000010DEsd000012A2bc03sc02i00 sudo -u $SUDO_USER curl -O https://download.open-mpi.org/release/open-mpi/v4.0/openmpi-4.0.3.tar.gz
# vendor : NVIDIA Corporation sudo -u $SUDO_USER tar xzf openmpi-4.0.3.tar.gz
# driver : nvidia-driver-418 - third-party non-free cd openmpi-4.0.3
# driver : nvidia-driver-450 - third-party non-free sudo -u $SUDO_USER ./configure --prefix=/usr/local/openmpi-4.0.3
# driver : nvidia-driver-460 - third-party non-free sudo -u $SUDO_USER make
# driver : nvidia-driver-450-server - distro non-free make install
# driver : nvidia-driver-460-server - distro non-free add_env PATH /usr/local/openmpi-4.0.3/bin
# driver : nvidia-driver-440 - third-party non-free add_env LD_LIBRARY_PATH /usr/local/openmpi-4.0.3/lib
# driver : nvidia-driver-418-server - distro non-free fi
# driver : nvidia-driver-465 - third-party non-free
# driver : nvidia-driver-470 - third-party non-free recommended #pick the latest one
# driver : nvidia-driver-410 - third-party non-free
# driver : nvidia-driver-470-server - distro non-free
# driver : nvidia-driver-455 - third-party non-free
# driver : xserver-xorg-video-nouveau - distro free builtin
# sudo apt-get install nvidia-driver-470 -y
# nvidia-smi # run this to check the driver is working
#root@ecs-testing:~# nvidia-smi
#Thu Dec 30 21:06:13 2021
#+-----------------------------------------------------------------------------+
#| NVIDIA-SMI 460.73.01 Driver Version: 460.73.01 CUDA Version: 11.2 |
#|-------------------------------+----------------------+----------------------+
#| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
#| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
#| | | MIG M. |
#|===============================+======================+======================|
#| 0 Tesla T4 Off | 00000000:21:01.0 Off | 0 |
#| N/A 61C P0 29W / 70W | 0MiB / 15109MiB | 0% Default |
#| | | N/A |
#+-------------------------------+----------------------+----------------------+
#
#+-----------------------------------------------------------------------------+
#| Processes: |
#| GPU GI CI PID Type Process name GPU Memory |
#| ID ID Usage |
#|=============================================================================|
#| No running processes found |
#+-----------------------------------------------------------------------------+
# install cuda/cudnn/nccl2 with apt-get arch=`uname -m`
# another option is to use linux.run https://developer.download.nvidia.com/compute/cuda/11.1.0/local_installers/cuda_11.1.0_455.23.05_linux.run sudo -u $SUDO_USER pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/${MINDSPORE_VERSION}/MindSpore/gpu/${arch}/${cuda_name}/mindspore_gpu-${version_map["$PYTHON_VERSION"]}-linux_${arch}.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple
sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
sudo add-apt-repository "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/ /"
sudo add-apt-repository "deb https://developer.download.nvidia.cn/compute/machine-learning/repos/ubuntu1804/x86_64/ /"
sudo apt-get update
sudo apt-get -y install cuda=${CUDA_VERSION}
# add cuda to path # check mindspore installation
cat >> ~/.bash_profile <<END
export PATH=/usr/local/cuda/bin:\$PATH
export LD_LIBRARY_PATH=\$LD_LIBRARY_PATH:/usr/local/cuda/lib64
END
source ~/.bash_profile
echo "cuda install success."
sudo apt-get install -y libcudnn8=${CUDA_VERSION} libcudnn8-dev=${CUDA_VERSION} libnccl2=${LIBNCCL2_VERSION} libnccl-dev=${LIBNCCL2_VERSION}
# optional (tensort for serving, openmpi for distributed training)
# uncomment this to compile openmpi
# cd /tmp
# curl -O https://download.open-mpi.org/release/open-mpi/v4.0/openmpi-4.0.3.tar.gz
# tar xvzf openmpi-4.0.3.tar.gz
# cd openmpi-4.0.3
# ./configure --prefix=/usr/local/openmpi
# make
# sudo make install
# echo 'export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/openmpi/lib' >> ~/.bash_profile
# echo 'export PATH=$PATH:/usr/local/openmpi/bin' >> ~/.bash_profile
# source ~/.bash_profile
# reference this to install tensorrt https://docs.nvidia.com/deeplearning/tensorrt/install-guide/index.html#downloading
echo "install mindspore gpu ${MINDSPORE_VERSION}"
pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/${MINDSPORE_VERSION}/MindSpore/gpu/${ARCH}/${CUDA_INSTALL_PATH}/mindspore_gpu-${version_map["$PYTHON_VERSION"]}-linux_${ARCH}.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple
# check if it is the right mindspore version
python -c "import mindspore;mindspore.run_check()" python -c "import mindspore;mindspore.run_check()"
# check if it can be run with GPU # check if it can be run with GPU
cd /tmp
cat > example.py <<END cat > example.py <<END
import numpy as np import numpy as np
from mindspore import Tensor from mindspore import Tensor
@ -130,5 +150,5 @@ x = Tensor(np.ones([1,3,3,4]).astype(np.float32))
y = Tensor(np.ones([1,3,3,4]).astype(np.float32)) y = Tensor(np.ones([1,3,3,4]).astype(np.float32))
print(ops.add(x, y)) print(ops.add(x, y))
END END
python example.py
python example.py cd -

View File

@ -0,0 +1,151 @@
#!/bin/bash
# Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================
# Prepare environment for mindspore gpu compilation on Ubuntu 18.04.
#
# This file will:
# - change deb source to huaweicloud mirror
# - install compile dependencies via apt like cmake, gcc
# - install python3 & pip3 via apt and set it to default
# - install CUDA by run file and cudnn via apt.
# - compile and install Open MPI if OPENMPI is set to on.
# - install LLVM if LLVM is set to on.
#
# Augments:
# - PYTHON_VERSION: python version to install. [3.7(default), 3.9]
# - CUDA_VERSION: CUDA version to install. [10.1(default), 11.1]
# - OPENMPI: whether to install optional package Open MPI for distributed training. [on, off(default)]
# - LLVM: whether to install optional dependency LLVM for graph kernel fusion. [on, off(default)]
#
# Usage:
# Need root permission to run, like `sudo bash -i ./ubuntu-gpu-source.sh`.
# To set augments, run it as `sudo PYTHON_VERSION=3.9 CUDA_VERSION=11.1 OPENMPI=on bash -i ./ubuntu-gpu-source.sh`.
set -e
PYTHON_VERSION=${PYTHON_VERSION:-3.7}
CUDA_VERSION=${CUDA_VERSION:-10.1}
OPENMPI=${OPENMPI:-off}
LLVM=${LLVM:-off}
available_py_version=(3.7 3.9)
if [[ " ${available_py_version[*]} " != *" $PYTHON_VERSION "* ]]; then
echo "PYTHON_VERSION is '$PYTHON_VERSION', but available versions are [${available_py_version[*]}]."
exit 1
fi
available_cuda_version=(10.1 11.1)
if [[ " ${available_cuda_version[*]} " != *" $CUDA_VERSION "* ]]; then
echo "CUDA_VERSION is '$CUDA_VERSION', but available versions are [${available_cuda_version[*]}]."
exit 1
fi
declare -A minimum_driver_version_map=()
minimum_driver_version_map["10.1"]="418.39"
minimum_driver_version_map["11.1"]="450.80.02"
driver_version=$(modinfo nvidia | grep ^version | awk '{printf $2}')
if [[ $driver_version < ${minimum_driver_version_map[$CUDA_VERSION]} ]]; then
echo "CUDA $CUDA_VERSION minimum required driver version is ${minimum_driver_version_map[$CUDA_VERSION]}, \
but current nvidia driver version is $driver_version, please upgrade your driver manually."
exit 1
fi
# add value to environment variable if value is not in it
add_env() {
local name=$1
if [[ ":${!name}:" != *":$2:"* ]]; then
echo -e "export $1=$2:\$$1" >> ~/.bashrc
fi
}
# use huaweicloud mirror in China
sed -i "s@http://.*archive.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
sed -i "s@http://.*security.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
# base packages
apt-get update
apt-get install software-properties-common lsb-release -y
apt-get install curl tcl automake autoconf libtool gcc-7 git libgmp-dev patch libnuma-dev flex -y
# cmake
wget -O - https://apt.kitware.com/keys/kitware-archive-latest.asc 2>/dev/null | apt-key add -
apt-add-repository "deb https://apt.kitware.com/ubuntu/ $(lsb_release -cs) main"
apt-get install cmake -y
# optional dependency LLVM for graph-computation fusion
if [[ X"$LLVM" == "Xon" ]]; then
wget -O - https://apt.llvm.org/llvm-snapshot.gpg.key | apt-key add -
add-apt-repository "deb http://apt.llvm.org/bionic/ llvm-toolchain-bionic-12 main"
apt-get update
apt-get install llvm-12-dev -y
fi
# optional openmpi for distributed training
if [[ X"$OPENMPI" == "Xon" ]]; then
origin_wd=$PWD
cd /tmp
sudo -u $SUDO_USER curl -O https://download.open-mpi.org/release/open-mpi/v4.0/openmpi-4.0.3.tar.gz
sudo -u $SUDO_USER tar xzf openmpi-4.0.3.tar.gz
cd openmpi-4.0.3
sudo -u $SUDO_USER ./configure --prefix=/usr/local/openmpi-4.0.3
sudo -u $SUDO_USER make
make install
add_env PATH /usr/local/openmpi-4.0.3/bin
add_env LD_LIBRARY_PATH /usr/local/openmpi-4.0.3/lib
cd $origin_wd
fi
# python
add-apt-repository -y ppa:deadsnakes/ppa
apt-get install python$PYTHON_VERSION python$PYTHON_VERSION-dev python$PYTHON_VERSION-distutils python3-pip -y
update-alternatives --install /usr/bin/python python /usr/bin/python$PYTHON_VERSION 100
# pip
sudo -u $SUDO_USER python -m pip install pip -i https://pypi.tuna.tsinghua.edu.cn/simple
update-alternatives --install /usr/bin/pip pip ~/.local/bin/pip$PYTHON_VERSION 100
update-alternatives --install /usr/local/bin/pip pip ~/.local/bin/pip$PYTHON_VERSION 100
sudo -u $SUDO_USER pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple
# install cuda/cudnn
cd /tmp
declare -A cuda_url_map=()
cuda_url_map["10.1"]=https://developer.download.nvidia.com/compute/cuda/10.1/Prod/local_installers/cuda_10.1.243_418.87.00_linux.run
cuda_url_map["11.1"]=https://developer.download.nvidia.com/compute/cuda/11.1.1/local_installers/cuda_11.1.1_455.32.00_linux.run
cuda_url=${cuda_url_map[$CUDA_VERSION]}
wget $cuda_url
sh ${cuda_url##*/} --silent --toolkit
cd -
apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
add-apt-repository "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/ /"
add-apt-repository "deb https://developer.download.nvidia.cn/compute/machine-learning/repos/ubuntu1804/x86_64/ /"
apt-get update
declare -A cudnn_name_map=()
cudnn_name_map["10.1"]="libcudnn7=7.6.5.32-1+cuda10.1 libcudnn7-dev=7.6.5.32-1+cuda10.1"
cudnn_name_map["11.1"]="libcudnn8=8.0.4.30-1+cuda11.1 libcudnn8-dev=8.0.4.30-1+cuda11.1"
apt-get install --no-install-recommends ${cudnn_name_map[$CUDA_VERSION]} -y
# add cuda to path
set +e && source ~/.bashrc
set -e
add_env PATH /usr/local/cuda/bin
add_env LD_LIBRARY_PATH /usr/local/cuda/lib64
add_env LD_LIBRARY_PATH /usr/lib/x86_64-linux-gnu
set +e && source ~/.bashrc
set -e
# wheel
sudo -u $SUDO_USER pip install wheel
# python 3.9 needs setuptools>44.0
sudo -u $SUDO_USER pip install -U setuptools
echo "The environment is ready to clone and compile mindspore."

View File

@ -1,44 +0,0 @@
#!/bin/bash
set -ex
# sudo cp -a /etc/apt/sources.list /etc/apt/sources.list.bak 单独执行
PYTHON_VERSION=${PYTHON_VERSION:-3.7.5}
MINDSPORE_VERSION=${MINDSPORE_VERSION:-1.5.0}
ARCH=`uname -m`
declare -A version_map=()
version_map["3.7.5"]="${MINDSPORE_VERSION}-cp37-cp37m"
#use huaweicloud mirror in China
sudo sed -i "s@http://.*archive.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
sudo sed -i "s@http://.*security.ubuntu.com@http://repo.huaweicloud.com@g" /etc/apt/sources.list
sudo apt-get update
# install python 3.7 and make it default
sudo apt-get install gcc-7 libgmp-dev curl python3.7 -y
sudo update-alternatives --install /usr/bin/python python /usr/bin/python3.7 100
cd /tmp
curl -O https://bootstrap.pypa.io/get-pip.py
sudo python get-pip.py
pip install https://ms-release.obs.cn-north-4.myhuaweicloud.com/${MINDSPORE_VERSION}/MindSpore/cpu/${ARCH}/mindspore-${version_map["$PYTHON_VERSION"]}-linux_${ARCH}.whl --trusted-host ms-release.obs.cn-north-4.myhuaweicloud.com -i https://pypi.tuna.tsinghua.edu.cn/simple
# check if it is the right mindspore version
python -c "import mindspore;mindspore.run_check()"
# check if it can be run with GPU
cat > example.py <<END
import numpy as np
from mindspore import Tensor
import mindspore.ops as ops
import mindspore.context as context
context.set_context(device_target="GPU")
x = Tensor(np.ones([1,3,3,4]).astype(np.float32))
y = Tensor(np.ones([1,3,3,4]).astype(np.float32))
print(ops.add(x, y))
END
python example.py