Skip to content

Latest commit

 

History

History
174 lines (109 loc) · 6.42 KB

MATLAB_installation.md

File metadata and controls

174 lines (109 loc) · 6.42 KB
Raw

Installation Instructions for MATLAB

Table of contents

Windows

Requirements:

  1. MATLAB
  2. Visual Studio (Community or Professional)
  3. A CUDA capable GPU from NVIDIA with compute capability greater or equal to 3.5
  4. CUDA Toolkit (9.2 or newer)

Tested on

Software Version
Windows 7, 8, 10.
MATLAB Any MATLAB >2016b
CUDA Any CUDA 9.2>
Visual Studio 2010 2013 2015 2019 2022

Simple Instructions

  1. Install MATLAB, Visual Studio and CUDA (Remember to install C++ when installing Visual Studio!)
  2. Run Compile.m

A successful installation should be able to execute the script at TIGRE/MATLAB/Demos/d03_generateData.m without errors.

Step by Step Instructions:

  1. Install MATLAB

  2. Install CUDA

    Any version above 9.2 has been tested, however its recommended to get the latests version as possible, for performance and support.
    CUDA download link
    Detailed installation guide
    NOTE: In windows at least, the User has to have no spaces.

  3. Install Visual Studio

    Make sure you install C++.
    Download link for the latest version
    Download link for older versions
    NOTE: In windows at least, the User has to have no spaces.

  4. Download TIGRE

    If you are using git, run: git clone https://github.com/CERN/TIGRE.git
    Manually download zip file otherwise.

  5. Test the correct configuration of Visual Studio

    Open MATLAB and run mex -setup -v. Among other things, the output should contain:

    ... Looking for compiler 'Microsoft Visual C++ 2015 (C)' ...
... Looking for registry setting 'HKLM\SOFTWARE\Microsoft\VisualStudio\SxS\VC7' 14.0 ...No.
... Looking for registry setting 'HKCU\SOFTWARE\Microsoft\VisualStudio\SxS\VC7' 14.0 ...No.
... Looking for registry setting 'HKLM\SOFTWARE\Wow6432Node\Microsoft\VisualStudio\SxS\VC7' 14.0 ...Yes ('C:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\').
... Looking for file 'C:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\bin\amd64\cl.exe' ...Yes.
... Looking for folder 'C:\Program Files (x86)\Microsoft Visual Studio 14.0\VC' ...Yes.

    Note the last 3 Yes. If those are not there it means that you do not have installed C++ on step 3 of this tutorial.
    To fix:
    -go to Control panel>Add or remove programs -> Visual studio community 20XX -> modify
    -Languages -> Visual C++

    Make sure that when you run mex -setup -v C++ is installed and Visual Studio is selected as the compiler for C/C++

  6. By opening MATLAB on yourTIGREpath/MATLAB, execute Compile.m

    If it fails, try opening mex_CUDA_win64.xml with your favorite editor and changing line 125 to link to your local nvcc location.

  7. Initialize TIGRE by typing InitTIGRE on the MATLAB Command Window.

  8. Run file TIGRE/MATLAB/Demos/d03_generateData.m. If it successfully executes, you have installed and compiled TIGRE properly.

If none of this works, please contact the authors at tigre.toolbox@gmail.com or ander.biguri@gmail.com

Linux

Requirements:

  1. MATLAB
  2. gcc
  3. A CUDA capable GPU from NVIDIA with compute capability greater or equal to 3.0
  4. CUDA Toolkit (9.2 or newer)

Tested on

Software Version
Ubuntu Any ubuntu 16.04>
MATLAB Any MATLAB 2016b>
CUDA Any Cuda 0.2>
gcc 6.4.0 7.2.0

Simple Instructions

  1. Install MATLAB, gcc and CUDA
  2. Run Compile.m

A successful installation should be able to execute the script at TIGRE/MATLAB/Demos/d03_generateData.m without errors.

Step by Step Instructions:

  1. Install MATLAB

  2. Install CUDA

    Installing CUDA in linux (specially one with a GUI) can be a challenge. Please follow NVIDIAs instructions carefully.
    CUDA download link

  3. Install gcc

    gcc should already be installed in your linux, as it is part of the linux distribution.
    If you need to install an older version of gcc, read here.

  4. Download TIGRE

    If you are using git, run: git clone https://github.com/CERN/TIGRE.git
    Manually download zip file otherwise.

  5. Make sure your terminal knows where CUDA is.

    For Ubuntu:

    • Using your favorite test editor, open ~/.bashrc. e.g. gedit ~/.bashrc.

    • Append to the file the following lines:

      export LD_LIBRARY_PATH=/usr/local/cuda/lib64
export PATH=$PATH:/usr/local/cuda/bin

    • restart your terminal

  6. From a terminal, execute MATLAB as matlab and run Compile.m located on TIGRE/MATLAB

  7. Run file TIGRE/MATLAB/Demos/d03_generateData.m. If it successfully executes, you have installed and compiled TIGRE properly.

If none of this works, please contact the authors at tigre.toolbox@gmail.com or ander.biguri@gmail.com

Advanced

If you are doing reconstruction of large datasets, and you want to use swap memory, you will need to deactivate TIGREs pinned memory feature at compile time. This will allow you to use swap memory, but it will make the operators in TIGRE slower, as pinned memory is used for simultaneous memory and compute.

You can do this by calling the Compile.m file from the MATLAB command line as Compile --no_pinned_memory.