How to compile:

For LINUX

    mkdir build

    cd build

    cmake ..

    make

For WINDOWS

    mkdir build

    cd build

    cmake .. -G "Visual Studio 12 Win64"

• Need to specify the compiler otherwise it will be defaulted to the 32 bit compiler, this example is assuming compiling in visual studio 2013

• click on the .sln

• compile (F7) on visual studio

Setup coding style

For Linux: Using Sublime 3 as main IDE

• Install prerequisites

    sudo apt-get install python (Ubuntu)
    sudo yum     install python (CentOS)

• Open Sublime

• Ctrl + `

• copy command from https://packagecontrol.io/installation

• press enter and restart Sublime

    sudo apt-get install python-pip (Ubuntu)
    sudo yum     install python-pip (CentOS)

    sudo pip install cpplint

    cd hvr-sample

    cp /usr/local/lib/python2.7/dist-packages/cpplint.py style-check/ (Ubuntu)
    cp /usr/lib/python2.7/site-packages/cpplint.py style-check/       (CentOS)

• (Copy cpplint.py to style-check when repo is cloned to local machine)

• Setup pre-push hook

    cp style-check/pre-push .git/hooks/pre-push

• Open Sublime

• ctrl + shift + p

• select "package control: install package"

• search and install SublimeLinter

• search and install SublimeLinter-cpplint

• add following to "Preferences"->"Package settings"->"SublimeLinter"->"Settings-User" under the cpplint section

 "cpplint": {
    "@disable": false,
    "args": [],
    "excludes": [],
    "filter": "-whitespace/braces,-whitespace/newline,-runtime/references,-build/c++11"
}

• If Settings-User is empty, "ctrl + shift + p" -> "SublimeLinter: enable"

• build the latest llvm/clang by following http://clang.llvm.org/get_started.html (obtain latest release)

• ctrl + shift + p

• select "package control: install package"

• search and install clang format

• "Preferences"->"Package settings"->"Clang Format"->"Settings-User"

• add in (To find "clang-format", use "updatedb -> locate clang-format")

{
    "binary": "/path/to/clang/bin/clang-format",

    "style": "Custom",

    "format_on_save": true
}

• add following style to "Preferences"->"Package settings"->"Clang Format"->"Custom Style-User"

{
    "BasedOnStyle": "Google",
    "AlignConsecutiveAssignments": true,
    "AlignTrailingComments": true,
    "AllowAllParametersOfDeclarationOnNextLine": true,
    "AllowShortCaseLabelsOnASingleLine": true,
    "AllowShortIfStatementsOnASingleLine": true,
    "AllowShortFunctionsOnASingleLine": "None",
    "AllowShortLoopsOnASingleLine": true,
    "BinPackArguments": false,
    "BinPackParameters": false,
    "BreakBeforeBraces": "Allman",
    "BreakConstructorInitializersBeforeComma": true,
    "ColumnLimit": 80
}

• ctrl + shift + p

• select "package control: install package"

• search and install SublimeLinter-contrib-cmakelint

• add following to "Preferences"->"Package settings"->"SublimeLinter"->"Settings-User" under the cmakelint section

"cmakelint": {
    "@disable": false,
    "args": [
        "--filter=-convention/filename"
    ],
    "excludes": []
}

• ctrl + shift + p

• search and install EditorConfig to get basic document consistency

For Windows: Using Visual Studio as main IDE

• Download and Install the latest python, add path for all users

    pip install cpplint

    copy C:\Python27\Scripts\cpplint-script.py style-check\cpplint.py

• Setup pre-push hook

    copy "style-check\pre-push" ".git\hooks\pre-push"  

• in Visual Studio, Tools -> External Tools -> Add. Name it cpplint, adjust the filter according to CPPLINT.cfg (this is in the repo)

    Title             : cpplint
    Command           : C:\Python27\python.exe
    Arguments         : C:\Python27\Scripts\cpplint-script.py --filter= --extensions=h,hpp,c,cpp $(ItemPath)
    Initial Directory : $(ItemDir)

• Move the newly created cpplint to the very top

• Download and install the latest clang-format plugin from http://llvm.org/builds/

• Go to Tools->Options->LLVM/Clang->ClangFormat and put down "File" in the Style option field.

• This will search for a .clang-format file in where the source is it's parent directory

• Install Visual Commander in Tools -> Extensions and Updates

• VCMD -> Extensions -> Add. Name it cpplint_clangformat. Choose C# as a language. Copy/Past the following code. Now Save, Compile, Install.

    using EnvDTE;

    using EnvDTE80;

     

    public class E : VisualCommanderExt.IExtension

    {

        public void SetSite(EnvDTE80.DTE2 DTE_, Microsoft.VisualStudio.Shell.Package package)

        {

            DTE = DTE_;
        
            events = DTE.Events;
        
            documentEvents = events.DocumentEvents;
        
            documentEvents.DocumentSaved += OnDocumentSaved;  

        }

     

        public void Close()

        {

            documentEvents.DocumentSaved -= OnDocumentSaved;

        }

     

        private void OnDocumentSaved(EnvDTE.Document doc)

        {

            if(doc.Language == "C/C++")

            {

                DTE.ExecuteCommand("Edit.SelectAll");

                DTE.ExecuteCommand("Tools.ClangFormat");

                DTE.ExecuteCommand("Tools.ExternalCommand1");

                DTE.ExecuteCommand("File.SaveAll");

            }

        }

     

        private EnvDTE80.DTE2 DTE;

        private EnvDTE.Events events;

        private EnvDTE.DocumentEvents documentEvents;

    }

• This will invoke cpplint and clang format whenever a c++ related file is saved

• in Visual Studio, Tools -> Extensions and Updates -> , search for editorconfig and install it

• in Visual Studio, Tools -> Extensions and Updates -> , search for cmake tools for visual studio and install it

Doxygen

• To build with doxygen

   cmake .. -DBUILD_DOCUMENTATION=ON 

For Linux

    sudo apt-get install doxygen texlive-full graphviz

For Windows

• Download and install the latest doxygen

• Download and install the latest MikTeX

• Download and install the latest Graphviz

• Manually add Graphviz to system PATH

CUDA

For Linux

• Download the latest cuda deb file on NVIDIA's website

    sudo apt-get remove --purge cuda*

    sudo dpkg -i cuda-repo-ubuntuXXXX_X.X-XX_amd64.deb

    sudo apt-get update

    sudo apt-get install cuda

• Add system path

    gedit ~/.bashrc

• export PATH=/usr/local/cuda/bin:$PATH

• export LD_LIBRARY_PATH=/usr/local/cuda/lib:$LD_LIBRARY_PATH

• export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH

• Copy and compile samples

    cuda-install-samples-X.X.sh ~

    cd NVIDIA_CUDA-X.X_Samples

    make

• Test samples

    cd bin/x86_64/linux/release
    
    ./deviceQuery

    ./bandwidthTest
    
    ./smokeParticles

For Windows

• Download the latest cuda

• Follow instructions on express installation

OpenCV

For Linux

• Install prerequisites

    sudo apt-get install cmake libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev

    sudo apt-get install python-dev python-numpy libtbb2 libtbb-dev libjpeg-dev libpng-dev libtiff-dev libjasper-dev libdc1394-22-dev

• Get the latest OpenCV

    git clone https://github.com/Itseez/opencv

    cd opencv
    
    mkdir build
    
    cd build

    ccmake ..

• Add/Remove modules accordingly

    make

    sudo make install

For Windows

• Download the latest OpenCV

• Set "OpenCV_DIR" system variable into "XXX\opencv\build"

• Set system variable at control panel->systems->advanced system settings->environment variables->system variables

Matlab runtime

For Linux

• Download the latest matlab runtime and unzip

    sudo ./install

For Windows

• Download the latest matlab runtime and install

GDB

Compile with GDB

• Before using GDB, any source file needs to be debugged must be compiled with -g flag

   g++ -g ...

Basic operations

• Start gdb

gdb executable_name 

• Set break point

b simple.cpp:147 (break on a line)
b main (break upon entering a function)
b simple.cpp:147 if x==0 (break on condition)

• get all break points information

b info

• Disable/enable a break point. (break point ID can be found with "b info")

disable 1 
enable 1

• Run the program

r (run)

• Continue running after stop

 c (continue)

• step. Execute next line.

n (next)

• Step in. Go inside a function call

s (step)

• Step out of a function. Jump out of current function.

finish 

• Show variable value (temporary)

print x
print x.count

• Call a funciton

print x.count()

• Display/undisplay variable (permanent)

display x
undisplay x

• Get the type of the variable

ptype x 

• Check memory.

x/256x ptr

This command shows 256 bytes presented as hexidecimal starting at address of ptr. Format can be changed to (d)ecimal, (u)nsigned, etc

• Quit gdb

quit

Other

• Find the source of "seg fault". At the most of the time, GDB will automatically stop at the place where the seg fault happened. This is a quick way to locate a seg fault.

gdb a.out
run

• cuda-gdb is capable of debugging cuda program running on the gpu and its usage is exactly the same as normal gdb.

cuda-gdb a.out