Micah P. Dombrowski / Aug 16 2018
with Dieter Komendera

CUDA Library Downloads

This downloads the Nvidia CUDA libraries, and compiles them all into an env for import into other articles. Note that each time, the actual download link must be updated by going to the linked address and loggin in with an Nvidia developer account, to get a working auth token.

1.
Downloads

1.1.
The CUDA Deep Neural Network library

cuDNN (Runtime and Dev).

FILENAME="libcudnn7_7.1.4.18-1+cuda8.0_amd64.deb"
wget -O /results/$FILENAME --progress=dot:giga \
  https://developer.download.nvidia.com/compute/machine-learning/cudnn/secure/v7.1.4/prod/8.0_20180516/Ubuntu16_04-x64/libcudnn7_7.1.4.18-1%2Bcuda8.0_amd64.deb?BJz_uwkj4kd2Orp-8uT9tL9xoSAQ11ObKqekJjLWqAw8lE0hv2OrOyKLL7CytI60YMGP_MwrWLgiN5nuYilppIpzWJsCCyTt0MIeeXX5Yv_9A3CH_T2UCLPiB_SGaW4wd0g3N1ALV5YCqxfT10FChQSWw2PE3qB1OL_YfUhSohOC7Qbzpcv_S7LzTbJzJJBakIjd1mHNYSbmrpKPbYZ5_6Ctq3BVMTbL6Lq0e8z1CneOLObpVk8

FILENAME="libcudnn7-dev_7.1.4.18-1+cuda8.0_amd64.deb"
wget -O /results/$FILENAME --progress=dot:giga \
  https://developer.download.nvidia.com/compute/machine-learning/cudnn/secure/v7.1.4/prod/8.0_20180516/Ubuntu16_04-x64/libcudnn7-dev_7.1.4.18-1%2Bcuda8.0_amd64.deb?wsjQUyVhV4_a7WExY3IbUvlxpomaJgytJv9zc2u1ihYCe-wjPue5yNCtXnfbq6hK_Cu996zONfZH7ffvN0ARlHpZToJIlwA05ttTLuCEDnA6mvAg5BN1jx2eCobxntuKaF8oURuofAK6wl1hxeFh3tG7Ucvbtamfe33RJkfOqR-ulXI9NNrHWWvK-ix-N41ieM6s5U1cjO0feiOfH60__XSZm_5y-MRNdAMJ3MG3eREi3L8o13nBFiPP
2.2s
libcudnn7-dev_7.1.4.18-1+cuda8.0_amd64.deb
libcudnn7_7.1.4.18-1+cuda8.0_amd64.deb

1.2.
NCCL

The Nvidia Collective Communications Library.

FILENAME="nccl-repo-ubuntu1604-2.2.13-ga-cuda8.0_1-1_amd64.deb"
wget -O /results/$FILENAME --progress=dot:giga \
  https://developer.download.nvidia.com/compute/machine-learning/nccl/secure/v2.2/prod/nccl-repo-ubuntu1604-2.2.13-ga-cuda8.0_1-1_amd64.deb?IjuL0-m989ZMYh8iuwQAvqi-b5-ljpdZF2LzczZyG1z2rKSmMOK2GKW842hX3BpWJ3wPEF2SxsEcdi4f5Fv7dTyvEZ6B0aeejI369OZaPmOpd-peMNATLISHiG7ALwyEbMEBkIsJIN9L_z7rmvXtsXzsPJiywQBF1uTCJ0FLo1Ftl3ublvr34JFzVMhaF0oqIZ88rEGK25OE9LOlAp-a9Kgc
0.8s
nccl-repo-ubuntu1604-2.2.13-ga-cuda8.0_1-1_amd64.deb

1.3.
TensorRT

FILENAME="nv-tensorrt-repo-ubuntu1604-cuda9.2-ga-trt4.0.1.6-20180612_1-1_amd64.deb"
wget -O /results/$FILENAME --progress=dot:giga \
  https://developer.download.nvidia.com/compute/machine-learning/tensorrt/secure/4.0/ga/nv-tensorrt-repo-ubuntu1604-cuda9.2-ga-trt4.0.1.6-20180612_1-1_amd64.deb?vqzwg-jItb_gujCx-CdzbzozoN7FoCBqzGg4GSjpIRFbfclefvr6n5S9TBSA-IIvjH2_-b-Ml4wvZrkJCnZ7dfaDZgIns7oG2V4_EzzMmSJStkJQS4QtWlpin-PY8FkO5mZQOXqAr2hyhzGi3YlZZDciDqX0509jIxwXCMroLIUWdiKYnxv-4IOYK4XNDWeugBVLcle6FlplRaSCvsWjR66S2fTEdRM_2oCB4SfxXCLpeKDH1n2I

FILENAME="TensorRT-4.0.1.6.Ubuntu-16.04.4.x86_64-gnu.cuda-9.2.cudnn7.1.tar.gz"
wget -O /results/$FILENAME --progress=dot:giga \
  https://developer.download.nvidia.com/compute/machine-learning/tensorrt/secure/4.0/ga/TensorRT-4.0.1.6.Ubuntu-16.04.4.x86_64-gnu.cuda-9.2.cudnn7.1.tar.gz?ufT_1eRISYwGEofM8H9i7K_QOrj408Yjlfq3Evk4cKsKxqJ9Awm3ejd7TB51TjDuWHShu2KCAmRfrVYMO9_K22XlJ81TP-nBwvv6TUI8X6ntafLBI-AjqGOF_JqvCDUDEgD7zvWv_od7qwoxgoWNweP016IxC3qBEmQslfUf7HFHfh4SR39zfuyUMqeeVjuVtqcZ3fOUtXJIV-ZZ8-okzUH84oLkSZETFiWABkOQe7d6vQ
41.2s
TensorRT-4.0.1.6.Ubuntu-16.04.4.x86_64-gnu.cuda-9.2.cudnn7.1.tar.gz
nv-tensorrt-repo-ubuntu1604-cuda9.2-ga-trt4.0.1.6-20180612_1-1_amd64.deb

2.
Compile & Test

cp libcudnn7_7.1.4.18-1+cuda8.0_amd64.debcuDNN↩ \
  /libcudnn7_7.1.4.18-1+cuda8.0_amd64.deb
cp libcudnn7-dev_7.1.4.18-1+cuda8.0_amd64.debcuDNN↩ \
  /libcudnn7-dev_7.1.4.18-1+cuda8.0_amd64.deb
cp nccl-repo-ubuntu1604-2.2.13-ga-cuda8.0_1-1_amd64.debNCCL↩ \
  /nccl-repo-ubuntu1604-2.2.13-ga-cuda8.0_1-1_amd64.deb
cp nv-tensorrt-repo-ubuntu1604-cuda9.2-ga-trt4.0.1.6-20180612_1-1_amd64.deb↩ \
  /nv-tensorrt-repo-ubuntu1604-cuda9.2-ga-trt4.0.1.6-20180612_1-1_amd64.deb
cp TensorRT-4.0.1.6.Ubuntu-16.04.4.x86_64-gnu.cuda-9.2.cudnn7.1.tar.gz↩ \
  /TensorRT-4.0.1.6.Ubuntu-16.04.4.x86_64-gnu.cuda-9.2.cudnn7.1.tar.gz
14.6s
dpkg -i *.deb
17.7s