Singularity images on Bridges

 

We have installed many of the NVIDIA GPU Cloud (NGC) containers as Singularity images on Bridges. These containers have been optimized for Volta and Pascal architectures by NVIDIA, including rigorous quality assurance.

These containers can all be found on Bridges under the directory /pylon5/containers/ e.g., /pylon5/containers/ngc/caffe. A link to further documentation is included below the table for each package when appropriate. The naming convention for most of these containers includes the year of the release as the first two digits and month as the next two, so release 18.10 was created in October 2018.

See also: 

 

More info: https://docs.nvidia.com/deeplearning/dgx/caffe-release-notes/index.html
Caffe
Available in /pylon5/containers/ngc/caffe
Caffe versionImage namePython versionOther supported software
0.17.3 caffe_20.03-py3.sif 3.6

OpenMPI 3.1.4
Tensorrt 7.0.0

0.17.1 18.10-py2.simg 2.7

OpenMPI 2.0

TensorRT 5.0.0 RC

0.17.1 18.09-py2.simg 2.7
0.17.1 18.08-py2.simg 2.7

OpenMPI 2.0

0.17.1 18.07-py2.simg 2.7

 

More info: https://docs.nvidia.com/deeplearning/dgx/caffe2-release-notes/index.html
Caffe2 version 0.81
Available in /pylon5/containers/ngc/caffe2
Image namePython versionOther supported software
18.08-py3.simg 3.5

OpenMPI 1.10.3

18.08-py2.simg 2.7
18.07-py3.simg 3.5
18.07-py2.simg 2.7

 

More info: https://docs.nvidia.com/deeplearning/dgx/cntk-release-notes/index.html
Microsoft Cognitive Toolkit (formerly CNTK) version 2.5
Available in /pylon5/containers/ngc/cntk
Image namePython versionOther supported software

18.08.py3.simg

3.6

  • OpenMPI 3.0.0
18.07.py3.simg

 

More info: https://docs.nvidia.com/deeplearning/digits/index.html
DIGITS 
Available in /pylon5/containers/ngc/digits
DIGITS versionImage namePython versionOther supported software
6.1.1 digits_20.03-tensorflow-py3.sif 3.6 Tensorrt 7.0.0
OpenMPI 3.1.4
NVIDIA CUDA 10.2.89
TensorFlow 1.15.2
6.1.1 digits_20.06-tensorflow-py3.sif 3.6 Tensorrt 7.1.2
OpenMPI 3.1.6
NVIDIA CUDA 11.0.167
TensorFlow 1.15.2
6.1.1 18.10.simg 2.7
  • NVCaffe 0.17.1
  • TensorFlow 1.10.0
  • TensorRT 5.0.0 RC
6.1.1 18.09.simg

 

Please note that the inferenceserver container is deprecated and will not be updated beyond 18.08.1. Starting with the 18.09 release, the TensorRT Inference Server is available in the tensorrtserver container.  See below.
More info: https://docs.nvidia.com/deeplearning/sdk/inference-release-notes
Inference Server
Available in /pylon5/containers/ngc/inferenceserver
Image namePython versionOther supported software
18.08.1-py3.simg  3.5  
18.08.1-py2.simg  2.7  
18.08-py3.simg  3.5
  • cuBLAS 9.0.425
  • cuDNN 7.2.1
  • NCCL 2.2.13
  • TensorRT 4.0.1
18.08-py2.simg  2.7

 

MATLAB
Available in /pylon5/containers/mdl
Image name matlab_r2019a.sif

 

More info: https://docs.nvidia.com/deeplearning/dgx/mxnet-release-notes/index.html
MXNet
Available in /pylon5/containers/ngc/mxnet
MXNET versionImage namePython versionOther supported software
1.6.0 mxnet_20.03-py3.sif 3.6 Tensorrt 7.0.0
OpenMPI 3.1.4
Horovod 0.19.1, NVIDIA Cuda 11.0.167
1.6.0 mxnet_20.06-py3.sif 3.6 Tensorrt 7.1.2
OPenMPI 3.1.6
Horovod 0.19.1, NVIDIA Cuda 11.0.167
1.3.0 18.10-py3.simg 3.5
  • ONNX exporter 0.1
  • Amazon Labs Sockeye sequence-to-sequence framework 1.18.28
  • TensorRT 5.0.0 RC
  • DALI 0.4 Beta
1.3.0 18.09-py3.simg 3.5
  • ONNX exporter 0.1
  • Amazon Labs Sockeye sequence-to-sequence framework 1.18.28
  • TensorRT 5.0.0 RC
  • DALI 0.2 Beta
1.2.0 18-08.py2.simg 2.7
  • ONNX exporter 0.1
  • Amazon Labs Sockeye sequence-to-sequence framework 1.18.28
  • TensorRT 4.0.1
  • DALI 0.1.2 Beta
1.2.0 18.07-py2.simg 2.7
  • ONNX exporter 0.1
  • Amazon Labs Sockeye sequence-to-sequence framework 1.18.23
  • TensorRT 4.0.1
  • DALI 0.1.1 Beta

 

More info: https://docs.nvidia.com/deeplearning/dgx/pytorch-release-notes/index.html
PyTorch 
Available in /pylon5/containers/ngc/pytorch
PyTorch versionImage namePython versionOther supported software
1.5.0a0+8f84ded pytorch_20.03-py3.sif 3.6 OpenMPI 3.1.4
Tensorrt 7.0.0
Apex
NVIDIA Cuda 10.2.89
Tensor Core optimized examples:
• ResNeXt101-32x4d
• SE-ResNext
• TransformerXL
• Jasper
• BERT
• Mask R-CNN
• Tacotron 2 and WaveGlow v1.1
• SSD300 v1.1
• Neural Collaborative Filtering (NCF)
• ResNet50 v1.5, GNMT v2
0.41+ 18.10-py3.simg
  • Caffe2
  • APEx
  • DALI 0.4 Beta
  • Tensor Core Optimized Examples:
    • ResNet50 v1.5
    • GNMT v2
  • TensorRT 5.0.0 RC
0.41+ 18.09-py3.simg
  • Caffe2
  • APEx
  • DALI 0.2 Beta
  • Tensor Core Optimized Examples:
    • ResNet50 v1.5
    • GNMT v2
  • TensorRT 5.0.0 RC

 

 

More info: https://docs.nvidia.com/deeplearning/dgx/tensorflow-release-notes/index.html
Tensorflow
Available in /pylon5/containers/ngc/tensorflow
Tensorflow versionImage namePython versionOther supported software
1.15.2 tensorflow_20.03-tf1-py3.sif 3.6 Horovod 0.19.0
OpenMPI 3.1.4
TensorRT 7.0.0
DALI 0.19.0
NVIDIA Cuda 10.2.89
Tensor Core optimized examples: (Included only in 20.03-tf1-py3)
• U-Net Medical
• SSD320 v1.2
• Neural Collaborative Filtering (NCF)
• BERT
• U-Net Industrial Defect Segmentation
• GNMT v2
• ResNet-50 v1.5
2.1.0  tensorflow_20.06-tf2-py3.sif
1.15.2 tensorflow_20.06-tf1-py3.sif
3.6 Horovod 0.19.1 for tf1 and Horovod 0.19.2 for tf2
OpenMPI 3.1.6
TensorRT 7.1.2
DALI 0.22
NVIDIA Cuda 11.0.167
Tensor Core optimized examples: (Included only in 20.06-tf1-py3)
• U-Net Medical
• SSD320 v1.2
• Neural Collaborative Filtering (NCF)
• BERT
• U-Net Industrial Defect Segmentation
• GNMT v2
• ResNet-50 v1.5
JupyterLab 1.2.2 including tensorboard
2.2.0 tensorflow_20.06-tf2-py3.sif
1.13.1 19.11-tf2-py3.simg 3.6
  • NVIDIA CUDA 10.2.89 (cuBLAS 10.2.2.89)
  • NVIDIA cuDNN 7.6.5
  • NVIDIA NCCL 2.5.6
  • Horovod 0.18.1
  • OpenMPI 3.1.4
  • TensorBoard 2.0.1
  • MLNX_OFED
  • TensorRT 6.0.1
  • DALI 0.15.0 Beta
  • Nsight Compute 2019.5.0
  • Nsight Systems 2019.5.2
  • Jupyter Client 5.3.4
  • Jupyter Core 4.6.1
  • Jupyter Notebook 6.0.1
  • JupyterLab 1.0.2
  • JupyterLab Server 1.0.0
  • Jupyter-TensorBoard
  tensorflow-19.05-py3.simg 3.5
  •  NVIDIA CUDA 10.1 Update 1 (cuBLAS 10.1 Update 1)
  • NVIDIA cuDNN 7.6.0
  • NVIDIA NCCL 2.4.6 
  • Horovod 0.16.1
  • OpenMPI 3.1.3
  • TensorBoard 1.13.1
  • MLNX_OFED 3.4
  • OpenSeq2Seq at commit 6e8835f
  • TensorRT 5.1.5
  • DALI 0.9.1 Beta
  • Nsight Compute 10.1.163
  • Nsight Systems 2019.3.1.94
  • Jupyter Client 5.2.4
  • Jupyter Core 4.4.0
  • JupyterLab 0.35.4
  • JupyterLab Server 0.2.0
  tensorflow-19.05-py2.simg 2.7
1.10.0 18.10-py3.simg 3.5

Horovod 0.13.10

OpenMPI 3.0.0

TensorBoard 1.10.0

MLNX_OFED 3.4

OpenSeq2Seq v18.10

TensorRT 5.0.0 RC

DALI 0.4 Beta

  18.10-py2.simg 2.7
  18.09-py3.simg 3.5

Horovod 0.13.10

OpenMPI 3.0.0

TensorBoard 1.10.0

MLNX_OFED 3.4

OpenSeq2Seq v18.09

TensorRT 5.0.0 RC

DALI 0.2 Beta

  18.09-py2.simg 2.7

 

 

TensorRT version 
Available in /pylon5/containers/ngc/tensorrt
TensorRT versionImage namePython versionOther supported software
7.0.0 tensorrt_20.03-py3.sif 3.6 OpenMPI 3.1.4
NVIDIA Cuda 10.2.89
7.1.2 tensorrt_20.03-py6.sif 3.6 OpenMPI 3.1.6
NVIDIA Cuda 11.0.167
5.0.0 RC 18.10-py3.simg 3.6
5.0.0 RC 18.10-py2.simg 2.7
5.0.0 RC 18.09-py3.simg 3.6
5.0.0 RC 18.09-p2.simg 2.7

 

More info: https://docs.nvidia.com/deeplearning/sdk/inference-release-notes/index.html
TensorRT Inference Server
Available in /pylon5/containers/ngc/tensorrtserver
TensorRT Inference Server versionImage namePython versionOther supported software
1.11.0 tensorrtserver_20.02-py3-clientsdk.sif 3.6 OpenMPI 3.1.4
TensorRT 7.0.0
0.7 Beta 18.10-py3.simg 3.5

OpenMPI 2.0

TensorRT 5.0.0 RC

0.6 Beta 18.09-py3.simg

 

More info: https://docs.nvidia.com/deeplearning/dgx/theano-release-notes/index.html
Theano version 1.02
Available in /pylon5/containers/ngc/theano
Image namePython versionOther supported software
18.08.simg

2.7

18.07.simg

 

More info: https://docs.nvidia.com/deeplearning/dgx/torch-release-notes/index.html
Torch version 7
Available in /pylon5/containers/ngc/torch
Image namePython versionOther supported software
18.08-py2.simg

2.7

18.07-py2.simg

Bridges and Bridges-AI Helping COVID-19 Research

Bridges and Bridges-AI offer exceptional capabilities for critical areas of research. Specific, unique strengths include AI, genomics, and collaborative workflows.

Read more about Bridges' capabilities for COVID-19 research.

New on Bridges

Increased time limit on RM and RM-shared

The maximum time limit has been increased to 72 hours on these partitions.
Read more

Omni-Path User Group

The Intel Omni-Path Architecture User Group is open to all interested users of Intel's Omni-Path technology.

More information on OPUG