change to use ubuntu22.04

This commit is contained in:
Your Name
2024-07-03 08:16:24 +00:00
parent 50f6dde35a
commit 6fcae83835

View File

@@ -5,7 +5,7 @@
# For reference:
# - https://docs.docker.com/build/dockerfile/frontend/#stable-channel
ARG CUDA_VERSION=12.1.0
FROM nvidia/cuda:${CUDA_VERSION}-cudnn8-devel-ubuntu20.04
FROM nvidia/cuda:${CUDA_VERSION}-cudnn8-devel-ubuntu22.04
ARG DEBIAN_FRONTEND="noninteractive"
ENV DEBIAN_FRONTEND=${DEBIAN_FRONTEND}
ENV MAMBA_ROOT_PREFIX=~/micromamba
@@ -85,7 +85,6 @@ ARG TORCHAUDIO_VERSION=2.3.1
ENV TORCHAUDIO_VERSION=${TORCHAUDIO_VERSION}
ARG PYTORCH_CUDA_VERSION=12.1
ENV PYTORCH_CUDA_VERSION=${PYTORCH_CUDA_VERSION}
ENV MLNX_OFED_VERSION=4.9-7.1.0.0
ARG SETUPTOOLS_VERSION=69.5.1
ENV SETUPTOOLS_VERSION=${SETUPTOOLS_VERSION}
ARG USE_CUDA=1
@@ -146,6 +145,7 @@ cd ..
rm -rf ${STAGE_DIR}/apex
EOT
ENV MLNX_OFED_VERSION=5.8-4.1.5.0
RUN <<EOT
#!/bin/bash
source /opt/conda/etc/profile.d/conda.sh
@@ -161,11 +161,11 @@ python3 -m pip uninstall -y deepspeed
# install deepspeed prepare
# install Mellanox OFED
mkdir -p ${STAGE_DIR}
wget -q -O - http://www.mellanox.com/downloads/ofed/MLNX_OFED-${MLNX_OFED_VERSION}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu20.04-x86_64.tgz | tar xzf -
cd MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu20.04-x86_64
wget -q -O - http://www.mellanox.com/downloads/ofed/MLNX_OFED-${MLNX_OFED_VERSION}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64.tgz | tar xzf -
cd MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64
./mlnxofedinstall --user-space-only --without-fw-update --all -q
cd ${STAGE_DIR}
rm -rf ${STAGE_DIR}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu20.04-x86_64*
rm -rf ${STAGE_DIR}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64*
cd ..
# install nv_peer_mem
rm -rf ${STAGE_DIR}
@@ -456,5 +456,5 @@ ARG NCCL_SOCKET_IFNAME='eth0'
# echo 'export PATH=/opt/conda/bin:$PATH' >> ~/.bashrc && \
# echo "source activate ${CONDA_ENV_NAME}" > ~/.bashrc
# CMD ["/usr/sbin/sshd", "-D"]
CMD ["/bin/bash", "-c", "/usr/sbin/sshd -D & while true; do sleep 1000; done"]
CMD ["/usr/sbin/sshd", "-D"]
# CMD ["/bin/bash", "-c", "/usr/sbin/sshd -D & while true; do sleep 1000; done"]