rocm4.0-pytorch1.7.0.dockerfile 4.03 KB
Newer Older
1
2
3
4
FROM rocm/pytorch:rocm4.0_ubuntu18.04_py3.6_pytorch_1.7.0

# OS:
#   - Ubuntu: 18.04
5
#   - OpenMPI: 4.0.5
6
#   - Docker Client: 20.10.8
7
8
9
10
11
12
13
# AMD:
#   - ROCm: 4.0
#   - HIP: 3.21.2
#   - MIOpen: 2.9.0
#   - RCCL: 2.7.8
# Mellanox:
#   - OFED: 5.2-2.2.3.0
14
15
# Intel:
#   - mlc: v3.9a
16
17
18
19

LABEL maintainer="SuperBench"

ENV DEBIAN_FRONTEND=noninteractive
20
RUN wget -qO - http://repo.radeon.com/rocm/rocm.gpg.key | APT_KEY_DONT_WARN_ON_DANGEROUS_USAGE=1 apt-key add - && \
21
22
23
24
25
26
27
28
29
30
    apt-get update && \
    apt-get install -y --no-install-recommends \
    autoconf \
    automake \
    build-essential \
    curl \
    dmidecode \
    git \
    jq \
    libaio-dev \
31
    libboost-program-options-dev \
32
    libcap2 \
33
    libnuma-dev \
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
    libpci-dev \
    libtinfo5 \
    libtool \
    lshw \
    net-tools \
    openssh-client \
    openssh-server \
    pciutils \
    util-linux \
    vim \
    wget \
    && \
    apt-get autoremove && \
    apt-get clean && \
    rm -rf /var/lib/apt/lists/* /tmp/*

50
51
52
53
54
55
56
# Install Docker
ENV DOCKER_VERSION=20.10.8
RUN cd /tmp && \
    wget https://download.docker.com/linux/static/stable/x86_64/docker-${DOCKER_VERSION}.tgz -O docker.tgz && \
    tar --extract --file docker.tgz --strip-components 1 --directory /usr/local/bin/ && \
    rm docker.tgz

57
58
59
60
61
62
63
64
65
66
# Update system config
RUN mkdir -p /root/.ssh && \
    touch /root/.ssh/authorized_keys && \
    mkdir -p /var/run/sshd && \
    sed -i "s/[# ]*PermitRootLogin prohibit-password/PermitRootLogin yes/" /etc/ssh/sshd_config && \
    sed -i "s/[# ]*PermitUserEnvironment no/PermitUserEnvironment yes/" /etc/ssh/sshd_config && \
    sed -i "s/[# ]*Port.*/Port 22/" /etc/ssh/sshd_config && \
    echo -e "* soft nofile 1048576\n* hard nofile 1048576" >> /etc/security/limits.conf && \
    echo -e "root soft nofile 1048576\nroot hard nofile 1048576" >> /etc/security/limits.conf

67
68
69
70
71
72
73
74
# Install OFED
ENV OFED_VERSION=5.2-2.2.3.0
RUN cd /tmp && \
    wget -q http://content.mellanox.com/ofed/MLNX_OFED-${OFED_VERSION}/MLNX_OFED_LINUX-${OFED_VERSION}-ubuntu18.04-x86_64.tgz && \
    tar xzf MLNX_OFED_LINUX-${OFED_VERSION}-ubuntu18.04-x86_64.tgz && \
    PATH=/usr/bin:${PATH} MLNX_OFED_LINUX-${OFED_VERSION}-ubuntu18.04-x86_64/mlnxofedinstall --user-space-only --without-fw-update --force --all && \
    rm -rf MLNX_OFED_LINUX-${OFED_VERSION}*

75
76
77
78
79
80
# Install OpenMPI
ENV OPENMPI_VERSION=4.0.5
RUN cd /tmp && \
    wget -q https://www.open-mpi.org/software/ompi/v4.0/downloads/openmpi-${OPENMPI_VERSION}.tar.gz && \
    tar xzf openmpi-${OPENMPI_VERSION}.tar.gz && \
    cd openmpi-${OPENMPI_VERSION} && \
81
    ./configure --enable-orterun-prefix-by-default --with-ucx=/usr --enable-mca-no-build=btl-uct && \
82
83
84
85
86
    make -j $(nproc) all && \
    make install && \
    ldconfig && \
    rm -rf /tmp/openmpi-${OPENMPI_VERSION}*

87
88
89
90
91
92
93
# Install HPC-X
RUN cd /opt && \
    wget -q https://azhpcstor.blob.core.windows.net/azhpc-images-store/hpcx-v2.8.3-gcc-MLNX_OFED_LINUX-${OFED_VERSION}-ubuntu18.04-x86_64.tbz && \
    tar xf hpcx-v2.8.3-gcc-MLNX_OFED_LINUX-${OFED_VERSION}-ubuntu18.04-x86_64.tbz && \
    ln -s hpcx-v2.8.3-gcc-MLNX_OFED_LINUX-${OFED_VERSION}-ubuntu18.04-x86_64 hpcx && \
    rm hpcx-v2.8.3-gcc-MLNX_OFED_LINUX-${OFED_VERSION}-ubuntu18.04-x86_64.tbz

94
95
# Install Intel MLC
RUN cd /tmp && \
96
97
    curl https://www.intel.com/content/dam/develop/external/us/en/documents/mlc_v3.9a.tgz -o mlc.tgz && \
    tar xzvf mlc.tgz Linux/mlc && \
98
    cp ./Linux/mlc /usr/local/bin/ && \
99
    rm -rf ./Linux mlc.tgz
100

101
102
ENV PATH="${PATH}" \
    LD_LIBRARY_PATH="/usr/local/lib:${LD_LIBRARY_PATH}" \
103
104
105
106
    SB_HOME=/opt/superbench \
    SB_MICRO_PATH=/opt/superbench \
    ANSIBLE_DEPRECATION_WARNINGS=FALSE \
    ANSIBLE_COLLECTIONS_PATH=/usr/share/ansible/collections
107
108
109
110

WORKDIR ${SB_HOME}

ADD third_party third_party
111
RUN ROCM_VERSION=rocm-4.0.0 make -C third_party -o rocm_rocblas rocm
112
113
114
115
116
117
118

# Workaround for image having package installed in user path
RUN mv /root/.local/bin/* /opt/conda/bin/ && \
    mv /root/.local/lib/python3.6/site-packages/* /opt/conda/lib/python3.6/ && \
    rm -rf /root/.local

ADD . .
119
RUN python3 -m pip install .[torch,ort] && \
120
121
    make cppbuild && \
    make postinstall