NVIDIA
NVIDIA
CUDA Deep Learning
Container
NVIDIA
NVIDIA
CUDA Deep Learning

CUDA is a parallel computing platform and programming model that enhances computing performance using NVIDIA GPUs. CUDA Deep Learning integrates networking and GPU-accelerated libraries like cuDNN, cuTensor, NCCL, HPC-x, and the CUDA Toolkit.

LayerLabelCreated
e79ee3fbedf892aa98ce3b9d0c309ad861a6277f7b3b3fbbe8734be6cc75f87aCONFIG
Entrypoint /opt/nvidia/nvidia_entrypoint.sh
12/05/2024 9:25 PM UTC
efb5bf568a93b4f3626e607babcf553a65618a15b59dcd3dcab16500ed9d4318COPY
/opt/amazon/aws-ofi-nccl /opt/amazon/aws-ofi-nccl
12/05/2024 9:25 PM UTC
e2387aa0ea719995aefc4ca1dea9a73f659482e91936fa1b10c68f59c7cca715ENV
LIBRARY_PATH=/usr/local/cuda/lib64/stubs:
12/05/2024 9:24 PM UTC
d26b5e2a627f7c012aac1cba0dd66d17d432e4946b172a78318d251f44118406RUN
GDRCOPY_VERSION=2.4.1 HPCX_VERSION=2.21 RDMACORE_VERSION=39.0 MOFED_VERSION=5.4-rdmacore39.0 OPENUCX_VERSION=1.18.0 OPENMPI_VERSION=4.1.7 EFA_VERSION=1.34.0 AWS_OFI_NCCL_VERSION=1.12.1 TARGETARCH=amd64 export DEVEL=1 BASE=0 &&
  /nvidia/build-scripts/installNCU.sh &&
  /nvidia/build-scripts/installCUDA.sh &&
  /nvidia/build-scripts/installLIBS.sh &&
  if [ ! -f /etc/ld.so.conf.d/nvidia-tegra.conf ]; then /nvidia/build-scripts/installNCCL.sh; fi &&
  /nvidia/build-scripts/installCUDNN.sh &&
  /nvidia/build-scripts/installCUTENSOR.sh &&
  /nvidia/build-scripts/installTRT.sh &&
  /nvidia/build-scripts/installNSYS.sh &&
  /nvidia/build-scripts/installCUSPARSELT.sh &&
  if [ -f "/tmp/cuda-${_CUDA_VERSION_MAJMIN}.patch" ]; then patch -p0 < /tmp/cuda-${_CUDA_VERSION_MAJMIN}.patch; fi &&
  rm -f /tmp/cuda-*.patch
12/05/2024 9:24 PM UTC
ab7d61b8a9559657624e677efc7a72de839047bb9b78639b8806eba326d9d7e9COPY
cuda-*.patch /tmp
12/05/2024 9:22 PM UTC
180f9db8d79b2173a03eeb7d2d7b629d71fa1f356fe56e62809eded39ecc3b98ENV
OMPI_MCA_coll_hcoll_enable=0
12/05/2024 9:22 PM UTC
a612985ee72d994427722ff0db561c2e608553e39a9bcce88ccc9be7fcf746daENV
OPAL_PREFIX=/opt/hpcx/ompi PATH=/usr/local/mpi/bin:/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/local/ucx/bin:/opt/amazon/efa/bin
12/05/2024 9:22 PM UTC
f931ce4a443b9257cd29ca4772085ce3cb8c933eece307fc3cca7e24ed113090RUN
GDRCOPY_VERSION=2.4.1 HPCX_VERSION=2.21 RDMACORE_VERSION=39.0 MOFED_VERSION=5.4-rdmacore39.0 OPENUCX_VERSION=1.18.0 OPENMPI_VERSION=4.1.7 EFA_VERSION=1.34.0 AWS_OFI_NCCL_VERSION=1.12.1 TARGETARCH=amd64 cd /nvidia &&
  ( export DEBIAN_FRONTEND=noninteractive &&
  apt-get update &&
  apt-get install -y --no-install-recommends libibverbs1 libibverbs-dev librdmacm1 librdmacm-dev libibumad3 libibumad-dev ibverbs-utils ibverbs-providers &&
  rm -rf /var/lib/apt/lists/* &&
  rm $(dpkg-query -L libibverbs-dev librdmacm-dev libibumad-dev | grep "\(\.so\|\.a\)$") ) &&
  ( cd opt/gdrcopy/ &&
  dpkg -i libgdrapi_*.deb ) &&
  ( cp -r opt/hpcx /opt/ &&
  cp etc/ld.so.conf.d/hpcx.conf /etc/ld.so.conf.d/ &&
  ln -sf /opt/hpcx/ompi /usr/local/mpi &&
  ln -sf /opt/hpcx/ucx /usr/local/ucx &&
  sed -i 's/^\(hwloc_base_binding_policy\) = core$/\1 = none/' /opt/hpcx/ompi/etc/openmpi-mca-params.conf &&
  sed -i 's/^\(btl = self\)$/#\1/' /opt/hpcx/ompi/etc/openmpi-mca-params.conf ) &&
  ( if [ ! -f /etc/ld.so.conf.d/nvidia-tegra.conf ]; then cd opt/amazon/efa/ &&
  dpkg -i libfabric*.deb &&
  rm /opt/amazon/efa/lib/libfabric.a &&
  echo "/opt/amazon/efa/lib" > /etc/ld.so.conf.d/efa.conf; fi ) &&
  ldconfig
12/05/2024 9:22 PM UTC
0eebc7efa1cb8cec1b48a227dd5681c168de606fea2b97531bc39ef108af92abARG
TARGETARCH=amd64
12/05/2024 9:22 PM UTC
86d41fb5c44814b5b1c214fd7c99333e86fd1cb8097e56e8b90d17195bd0e849ENV
GDRCOPY_VERSION=2.4.1 HPCX_VERSION=2.21 MOFED_VERSION=5.4-rdmacore39.0 OPENUCX_VERSION=1.18.0 OPENMPI_VERSION=4.1.7 RDMACORE_VERSION=39.0 EFA_VERSION=1.34.0 AWS_OFI_NCCL_VERSION=1.12.1
12/05/2024 9:22 PM UTC