document.tex

\documentclass[11pt]{book}
\usepackage{draftcopy} % PS or DVI
% \usepackage{pdfdraftcopy} % PDF
% http://filoxus.blogspot.com/2008/01/how-to-insert-watermark-in-latex.html
\usepackage{amsmath} %Never write a paper without using amsmath for its many new commands

\usepackage{amssymb} %Some extra symbols

\usepackage{makeidx} %If you want to generate an index automatically
%\usepackage{graphicx} %If you want to include postscript graphics
\ifx\pdftexversion\undefined
\usepackage[dvips]{graphicx}
%%%%%%%% 
%% To include .jpg and .pnm into latex
%%%%%%%%%%%%%%
\DeclareGraphicsExtensions{.jpg,.eps,.pnm}
\DeclareGraphicsRule{.jpg}{eps}{.jpg.bb}{`jpeg2ps -h #1}
\DeclareGraphicsRule{.pnm}{eps}{.pnm.bb}{`pnmtops #1}
\else
\usepackage{graphicx}
\fi 
  
\usepackage{epsfig}

\usepackage{color}  %Add color capabilities to fonts,e.g. \color{blue}{text something}
\usepackage{listings} % Add typeset programs (programming code) within LaTeX.

\usepackage{tabularx}
%\usepackage{ucs}
%\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{lmodern}

\usepackage{chemarr}

\usepackage{verbatim}

\usepackage[version=3]{mhchem}
\usepackage[style=2]{mdframed}

\usepackage{mathtools}
\DeclarePairedDelimiter\ceil{\lceil}{\rceil}
\DeclarePairedDelimiter\floor{\lfloor}{\rfloor}

%\usepackage{mystyle} 
%Create your own file, mystyle.sty where you put 
% all your own \newcommand statements, for example.

%\usepackage{url}
\usepackage[hyphens]{url}
           % usage \url{URL}

%\usepackage[hypertex, dvips]{hyperref}       % does not work with LaTeX2HTML
            % usage \href{URL}{text}
\usepackage[breaklinks=true, linktocpage]{hyperref}              
            
            
%\includeonly{chaptr2} %If you just want to process chaptr2.tex

\usepackage{breakurl}  % used with dvips, not with pdflatex
\usepackage{breakcites}

\usepackage[all]{hypcap}
%\graphicspath{{./images/}}

\usepackage{lgrind}

\usepackage[stable]{footmisc}

\usepackage{multicol}

\usepackage{fancyvrb}

\usepackage{ulem} 
% support : \uline, \sout, \uuline...
\usepackage{soul}
\usepackage{framed}
\usepackage[numbers]{natbib}
\usepackage{multicol} % for layout multicolumn on pages
\usepackage{multirow} % for table spanning multicolumn, multirow
%
% http://andrewjpage.com/index.php?/archives/43-Multirow-and-multicolumn-spanning-with-latex-tables.html

\usepackage{import} 


%%% definition %%%%%%%%%%%
\def\mus{{\mu\text{s}}}


\begin{document}

\author{Tuan, Hoang-Trong}
\title{GPU Computing}
\date{Jan 2009}

\frontmatter
\tableofcontents
%\include{preface}

\mainmatter
% \lstset{language={[77]Fortran}, numbers=left, numberstyle=\tiny,
%   stepnumber = 5, numbersep=5pt, keywordstyle=\color{blue}}

\part{Part I: Architecture}

\include{GPGPU_Architectures}
\import*{../Reverse_Engineering/}{Library.tex}
\include{PTX_lang}
\import*{../Android/}{SoC.tex}
\include{Tegra_SoC}
\import*{../Android/}{AppleA.tex}
\include{Linux_kernel_support}
 
\part{Part II: Language/Extension}

\lstset{language={C}, numbers=left, numberstyle=\tiny,
  stepnumber = 5, numbersep=5pt, keywordstyle=\color{blue}}
\include{Graphics_Languages}
\include{CCUDA}
\include{CUDAdriverAPI}
\include{FCUDA}
\include{CUDA_Cpp}
\include{CUDA_CSharp}

\include{OpenCL}

\import*{../C-Cpp_Manual/}{Compile.tex}


\part{Part III: Code abstract Level}

\lstset{language={[77]Fortran}, numbers=left, numberstyle=\tiny,
  stepnumber = 5, numbersep=5pt, keywordstyle=\color{blue}}
\include{AcceProgModel}
\include{CUDAFortranC}
\include{CUDA_Windows}
\include{OpenACC}

\part{Part IV: GPU-GPU, node-node}
\include{Multiple_GPUs}
\include{CUDA_Samples}
\import*{../Parallel_Programming/}{MPI_Extension.tex}


\part{Part V: Graphics}
 
\include{CUDA_Graphics}
\include{VSCSE}
\include{DirectCompute}
\include{Video_codecs}

\part{Part VI: Benchmark/Profile}
%\part{Benchmarking \& Tools}
\include{Timing_CUDAcode}
\include{Debug_Tools}
\include{ProfilingTool}
\include{Benchmark_Leak}
\include{Benchmark_ECC}
\include{TroubleShooting}

\part{Part VII: Packages for CUDA}
\import*{../Probability/}{Differentiation.tex}
\import*{../Fortran_Manual/}{Integration.tex}
\import*{../Fortran_Manual/}{ODE.tex}
\import*{../Fortran_Manual/}{PDE.tex}
\include{cuBLAS}
\include{cuFFT}
\include{Thrust}
\include{cuDPP}
\include{CULA}
\include{cuSP}
\include{NPP}
\include{matCUDA}
\include{cuRAND}
\include{Package4CUDA}


% 
% \part{Applications}
% \include{MedicalPhysics}
% \include{UltraSound}
% \include{MicroEmboli_Detection}

%\part{Appendix}
%\include{Appendix}
%\include{CUDAonWindows}
Convention:

\begin{itemize}
\item Text with \textcolor{red}{red color}: obsolete feature,
  statement...
\item Text with \textcolor{green}{yellow color}: not available in this version
\end{itemize}
\backmatter 
%\include{glossary}
%\include{notat}
\bibliographystyle{amsalpha} %The style you want to use for references.
\bibliography{../ref_Fortran} %The files containing all
                                %the articles and books you ever
                                %referenced. 
\printindex %Make an index AUTOMATICALLY

\end{document} 

% The share memory maps to the
% scratchpad memory of the GPU, and is local to each thread block.

% The texture memory use GPU texture caching and filtering capabilities,
% best utilized with data access patterns exhibiting 2D locality (how
% about 3D????)

% Logical threads can be divergent via a stack-based reconvergence
% algorithm with masked execution[3]