Torch Release 18.01

The NVIDIA container image of Torch, release 18.01, is available.

Torch container image version 18.01 is based on Torch7.

Contents of Torch

This container image contains the complete source of the version of NVIDIA Torch in /opt/torch. It is pre-built and installed into the /usr/local/[bin,share,lib] directories in the container image.

The NVIDIA^® Collective Communications Library ™ (NCCL) (NCCL) library and Torch bindings for NCCL are installed in this container, and models using DataParallelTable can easily leverage this library for fast parallel training. The container also includes the following:

• Ubuntu 16.04 including Python 2.7 environment
• NVIDIA CUDA 9.0.176 including CUDA^® Basic Linear Algebra Subroutines library™ (cuBLAS) 9.0.282
• NVIDIA CUDA^® Deep Neural Network library™ (cuDNN) 7.0.5
• NCCL 2.1.2 (optimized for NVLink™ )

Driver Requirements

Release 18.01 is based on CUDA 9, which requires NVIDIA Driver release 384.xx.

Key Features and Enhancements

This Torch release includes the following key features and enhancements.

• Latest version of cuBLAS
• Latest version of cuDNN
• Latest version of NCCL
• Ubuntu 16.04 with December 2017 updates

Known Issues

cuBLAS 9.0.282 regresses RNN seq2seq FP16 performance for a small subset of input sizes. As a workaround, revert back to the 11.12 container.