Debian Codegraph @

doc/Latex/apps/lrccs/LRCCS_Manual.tex

raw
1 \documentclass[10pt,a4paper]{scrartcl}
2
3 \usepackage{amsmath}
4 \usepackage{amsfonts}
5 \usepackage{amssymb}
6 \usepackage{color}
7 \usepackage[english]{babel}
8 \usepackage[
9 colorlinks=true,
10 urlcolor=blue,
11 linkcolor=green
12 ]{hyperref}
13
14 % no indentation
15 \setlength{\parindent}{0pt}
16
17 \begin{document}
18 \section*{MADNESS: CIS Excitation Energies}
19 The Tutorial was written for the Git commit \textit{f03a51fbe4f5a30d870c28772e5960ae39502da9} which you can checkout before compiling via: \\
20 \textit{git checkout f03a51fbe4f5a30d870c28772e5960ae39502da9}\\
21 If you experience any sort of trouble please feel free to contact me at:\\
22 \href{mailto:[email protected]}{[email protected]}
23
24 \section{Built LRCCS}
25 Detailed information can be found in the MADNESS wiki on Github: \url{https://github.com/m-a-d-n-e-s-s/madness/wiki}
26 \begin{itemize}
27 \item Make shure you have Git installed
28 \item Make shure you have (see also the MADNESS wiki):
29 \begin{itemize}
30 \item C++11 Compiler: Clang-3.8 or later (Mac OS), GCC-4.8 or later (Linux)
31 \item Corresponding MPICH
32 \item GNU Autotools (autoconf 2.59 or later, automake 1.11 or later)
33 \end{itemize}
34 \item Get the MADNESS source Code from Github\\
35 \textit{git clone {https://github.com/m-a-d-n-e-s-s/madness.git} madness-source}
36 \item Execute \textit{autogen.sh} in the madness-source directory
37 \item Configure MADNESSS in your compile directory like this (use the correct paths):\\
38 \textit{/path-to-source/madness-source/configure 'CC=/usr/bin/clang' 'CXX=/usr/bin/clang++'
39 'MPICC=/opt/local/bin/mpicc-mpich-clang38' 'MPICXX=/opt/local/bin/mpicxx-mpich-clang38'}
40 \item Compile the MADNESS libraries in your compile directory\\
41 \textit{make -C /path-to-madness-compile/madness-compile/ libraries}
42 \item Compile cis in the MADNESS examples directory\\
43 \textit{make -C /path-to-madness-compile/madness-compile/src/apps/cis cis}
44 \end{itemize}
45
46 \section{Run LRCCS}
47 After compiling you can execute \textit{cis} in the \textit{madness-compile/src/apps/cis} directory. For the calculation it is necessary that an input file with the name \textit{input} is present in the directory from which you call \textit{cis}.
48
49 \clearpage
50 \section{Quick Tutorial}
51 The programm \textit{LRCCS} \cite{mra-cis} will solve the linear-response equations for excitation energies in the CIS approximation via the following steps.
52 \begin{enumerate}
53 \item Solve Hartree-Fock (HF) ground state
54 \item Create guess vectorfunctions $\lbrace x_i \rbrace $ from the HF orbitals $\lbrace \phi_i \rbrace$ via
55 \[ x_i = f\cdot\phi_i \]
56 where $f$ is constructed from polynomials (see Keyword section)
57 \item Iterate the guess vectorfunctions
58 \item Iterate the lowest guess vectorfunctions further
59 \item Iterate sequentially untill convergence
60 \end{enumerate}
61 Detailed information can be found in \cite{mra-cis}.
62 \subsection{Input File Structure}
63 The input file has four main sections.
64 \paragraph{Ground-State section:} Begins with \textit{dft} and ends with \textit{end}\\
65
66 Mandatory Keywords:
67 \begin{itemize}
68 \item xc value: \textit{value} is the exchange correlation potential. To solve the CIS equations this has to be always \textit{hf}
69 \item k value: \textit{value} is the order of polynomials for the MRA representation (usually around 6-9)
70 \item econv value: \textit{value} is the MRA:threshold for the HF orbitals and at the same time the convergence threshold for the energy
71 \end{itemize}
72 Optional Keywords:
73 \begin{itemize}
74 \item L value: \textit{value} it the size of the cubic simulation box in atomic units. Each dimension of the box runs rum -L to L (default is 50.0)
75 \item dconv value: \textit{value} convergence threshold for the orbitals (does not affect MRA threshold)
76 \item nuclear\_corrfac value: \textit{value} is the nuclear correlation factor which is used \\
77 (default is \textit{none}, other options are \textit{slater}, \textit{GradientalGaussSlater}, \textit{GaussSlater}, \textit{LinearSlater} and
78 \textit{Polynomial}). For more information see \cite{nemo-I}.
79 \item no\_compute or restart: Load the HF equations from the file \textit{restartdata.00000} which has to be present in the same directory. If \textit{restart} is chosen the HF orbitals will be re-iterated
80 \end{itemize}
81
82 \paragraph{Response Section:} Begins with \textit{cc2} and ends with \textit{end}\\
83
84 Mandatory Keywords
85 \begin{itemize}
86 \item thresh\_3D value: \textit{value} is the MRA threshold for the response vectorfunctions (default is 5)
87 \end{itemize}
88 Optional Keywords
89 \begin{itemize}
90 \item tda\_econv\_hard value: \textit{value} is the final convergence threshold for the excitation energies
91 \item tda\_dconv\_hard value: \textit{value} is the final convergence threshold for the response vectorfunctions
92 \item tda\_excitations value: \textit{value} is the number of excitation energies which shall be calculated
93 \item tda\_guess\_excitations value: \textit{value} is the number of guess excitations which is iterated in the beginning
94 \item tda\_iterating\_excitations value: \textit{value} is the number of excitation vectors which are iterated together (change if you have memory problems)
95 \item freeze value: \textit{value} is the number of frozen core orbitals
96 \item tda\_guess value: \textit{value} is the guess that is used (default is a big guess from a perturbed fock matrix). Possible entries are: \textit{dipole}, \textit{dipole+}, \textit{quadrupole}, \textit{big\_fock\_3}, \textit{big\_fock\_4}, \textit{c2v}, \textit{c2v\_big} and \textit{custom}).
97 The polynomials which are used for the specific guesses will be displayed in the output. The \textit{custom} valueword has to be used together with the \textit{exop} valueword.
98 \item exop value: \textit{value} is a custom polynomial to create a guess vectorfunction from the HF orbitals.
99 Example: \textit{exop x 1.0 y 2.0, x 3.0 z 1.0 c -2.0} corresponds to $xy^2 + -2.0\cdot x^3z$.
100 To calculate more than one guess the exop keyword can used more than one time.
101 Example: The following line would be equivalent with the \textit{tda\_guess dipole+} keyword: \\
102 \textit{tda\_guess custom} \\
103 \textit{exop x 1.0}\\
104 \textit{exop y 1.0}\\
105 \textit{exop z 1.0}\\
106 \textit{exop x 2.0, y 2.0, z 2.0}
107 \end{itemize}
108 \clearpage
109 \paragraph{Geometry section:} Begins with \textit{geometry} and ends with \textit{end}. Contains the molecular coordinates in atomic units
110 \paragraph{Plot section:} Begins with \textit{plot} and ends with \textit{end}. Contains plotting information. Can in principle be empty but should be present.
111 \subsection{Example Input}
112 The example input File should also be found in the madness-source/src/apps/cis directory with the name \textit{input}.\\
113
114
115 dft\\
116 xc hf\\
117 econv 1.e-5\\
118 dconv 1.e-4\\
119 end\\
120
121 cc2\\
122 thresh\_3D 1.e-4\\
123 tda\_econv\_hard 1.e-4\\
124 tda\_dconv\_hard 1.e-3\\
125 tda\_guess dipole+ \\
126 end\\
127
128 geometry\\
129 he 0.0 0.0 0.0\\
130 end\\
131
132 plot \\
133 plane x1 x2\\
134 origin 0.0 0.0 0.0\\
135 zoom 1.0\\
136 end
137
138 %References
139 \bibliographystyle{plain}
140 \bibliography{references}
141
142 \end{document}