Installation

Quick Start

  1. Download the appropriate NVidia driver from http://www.nvidia.com/download/index.aspx See available drivers in Supported NVIDIA Drivers.

    Note

    The driver .run file will be named NVIDIA-Linux-x86_64-352.*.run We strongly advise manually downloading and installing the appropriate NVidia driver for your system as opposed to using a package manager.

  2. Once you have downloaded the driver, change to the directory containing the driver package and install the driver by running, as root, sh ./NVIDIA-Linux-x86_64-***.**.run See http://docs.nvidia.com/cuda/cuda-installation-guide-linux/#runfilei for more detailed installation instructions.

    Note

    FastROCS TK does not require the CUDA Toolkit/runtime/SDK to be installed. Only the NVidia driver need be installed using the .run file named NVIDIA-Linux-x86_64-***.*.run Please ensure you install a supported driver.

    Note

    The output of the nvidia-smi command is extremely useful when debugging FastROCS TK issues. Please include the output from nvidia-smi in any request to support@eyesopen.com.

    Warning

    The NVidia kernel module can often conflict with the open source Nouveau display drivers depending on your specific Linux distribution. The NVidia documentation is a much more complete and up-to-date source for information on how to work around this issue. See: http://docs.nvidia.com/cuda/cuda-installation-guide-linux/#runfile-nouveau

  3. Install the OpenEye Toolkits into a virtual environment

    $ mkvirtualenv fastrocs
    (fastrocs) $ pip install --extra-index-url https://pypi.anaconda.org/OpenEye/simple OpenEye-toolkits
    
  4. Once installed, follow the instructions in Tutorial 0: Basic usage of FastROCS TK to get you started.

Supported Hardware

FastROCS TK is currently only supported on Linux with NVidia graphics cards. FastROCS TK will work with any NVidia GPU released with a compute capability of 3.0 or higher. This includes graphics cards for high performance computing (Tesla), professional workstations (Quadro), and gaming (GeForce). For a comprehensive table of which GPUs fall into which compute capability category please refer to the CUDA wikipedia page. FastROCS TK has dropped support for all Nvidia Tesla, GEForce and Quadro cards with a compute capability < 3.0. This includes all GPUs with Fermi architecture, i.e. Tesla C2050 & GEForce GTX 430. The following graph shows the performance of FastROCS TK on some modern GPU models.

../_images/FastROCS_Latest_GPU_Comparison.png

Note

Performance can vary with driver versions and other factors. The above graph is meant as a general guide of what performance customers can expect. It is highly recommended to go with a certified vendor like Exxact to ensure best performance.

Supported Operating Systems

  • RHEL6 / CentOS 6
  • RHEL7 / CentOS 7
  • Ubuntu 14.04
  • Ubuntu 16.04
  • Ubuntu 18.04

Supported NVIDIA Drivers

FastROCS TK Release NVIDIA Driver
1.9.0 (2018.Oct) 361.*, 367.*, 375.*, 381.*, 384.*, 387.*, 390.*, 396.*
1.8.3 (2017.Oct) 346.*, 352.*, 361.*, 367.*, 375.*, 381.*, 384.*, 387.*
1.8.2 (2017.Jun) 346.*, 352.*, 361.*, 367.*, 375.*, 381.*
1.8.1 (2017.Feb) 346.*, 352.*, 361.*, 367.*, 375.*
1.8.0 (2016.Oct) 346.*, 352.*, 361.*, 367.*
1.7.0 (2016.Jun) 340.*, 346.*, 352.*, 361.*
1.6.0 (2016.Feb) 340.*, 346.*, 352.*
1.5.1 (2015.Oct) 352.*
1.5.0 (2015.Jun) 346.*