Linux

The latest TiViPE release for Linux Ubuntu 18.04 is 2.1.4 (October 2018).

TiViPE214Ubuntu1804.tgz
[688 MB]

TiViPE can be downloaded for free under GNU-GPL license. Commercial developers however need a different license, please contact us. Download TiViPE214Ubuntu1804.tgz  for Linux Ubuntu 18.04.

Open a terminal (ctrl-alt-t) and copy the downloaded file to the home directory, and extract its data:

cd $HOME
tar xzf TiViPE214Ubuntu1804.tgz

Open your favorite editor and add

# --> Read $HOME/TiViPE/vars, if present.
if [ -f $HOME/TiViPE/vars ]; then
 . $HOME/TiViPE/vars 
fi

to the $HOME/.bashrc file.

Open a new terminal and type

TiViPE

on the command line to start TiViPE.

Validate that all the packages required as described below are installed. Check the file $HOME/TiViPE/vars and verify that all paths and variables have been set properly.

Packages that need to be installed in the Ubuntu distribution

TiViPE relies on a number of programs. By default gcc, tar, perl, and ghostscript have been installed in the Ubuntu distribution. In addition the following packages are needed:

  • doxygen
  • fftw3
  • ffmpeg
  • texlive-full
  • imagemagick
  • libinsighttoolkit4-devlibgl1-mesa-dev
  • graphviz
  • putty
  • libgtk2.0-dev
  • pkg-config

The packages are installed by typing sudo apt install followed by the name of the package.

Qt

TiViPE makes use of Qt, it is downloaded from

https://www.qt.io/download-thank-you?os=linux (qt-unified-linux-x64-3.0.5-online.run)

Install qt:

chmod 755 unified-linux-x64-3.0.5-online.run
sudo ./qt-unified-linux-x64-3.0.5-online.run

Select the 5.11.2 (or newer) gcc 64 bit version and the source code. Qt will be installed in /opt/Qt.
The directory installed should be the same as the QTDIR variable in $HOME/TiViPE/vars.

NVIDIA CUDA

Download cuda_10.0.130_410.48_linux.run from

https://developer.nvidia.com (select: Linux, x86_64, Ubuntu, 18.04, runfile)

and install cuda

chmod 755 cuda-10.0.130_410.48_linux.run
sudo ./cuda-10.0.130_410.48_linux.run

Next take the following steps

accept install driver 410.48 (yes; no if there is no NVIDIA hardware)
opengl libs (yes)
run nvidia-xconfig (no)
cuda (yes)
link (yes)

Cuda is installed at /usr/local/cuda-10.0 and a symbolic link is created to /usr/local/cuda.

Deep learning support using CUDA requires the cuDNN package:

https://developer.nvidia.com (cudnn-10.0-linux-x64-v7.3.0.29.tgz)

Unzip the package:

tar -xzvf cudnn-10.0-linux-x64-v7.3.0.29.tgz

and install the files:

sudo cp cuda/include/cudnn.h /usr/local/cuda/include
sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64
sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*