scnc
/
qmcpack
mirror of https://github.com/QMCPACK/qmcpack.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

htxelatex is more picky than xelatex

This commit is contained in:
Peter Doak 2018-06-09 08:55:35 -04:00
parent 5a66ae6c93
commit 0d6106e10b
9 changed files with 59 additions and 77 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

64
manual/analysis.tex
View File

@ -116,11 +116,10 @@ by plotting the trace for each QMC run.
\begin{figure}
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.75\columnwidth]{figures/qmca_mean_error_trace.pdf}
\end{center}
\caption{Trace of the VMC local energy for an 8 atom cell of diamond generated with \texttt{qmca}. The x-axis (``samples'') refers to the VMC block index in this case.
\includegraphics[trim = 0mm 0mm 0mm 0mm,clip,width=0.75\columnwidth]{figures/qmca_mean_error_trace.pdf}
\caption{Trace of the VMC local energy for an 8 atom cell of diamond generated with \texttt{qmca}. The x-axis (``samples'') refers to the VMC block index in this case.}
\label{fig:qmca_mean_error_trace}
}
\end{center}
\end{figure}
If we exclude none of the equilibration data points, we get an
@ -206,8 +205,8 @@ proceed without using the ``\texttt{-e}'' option with \texttt{qmca}.
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.9\columnwidth]{figures/qmca_judge_opt.pdf}
\end{center}
\caption{Trace of the local energy during one- and two-body Jastrow optimization for an 8 atom cell of diamond generated with \texttt{qmca}. Data for each optimization cycle (QMCPACK series) is separated by a vertical black line.
\label{fig:qmca_judge_opt}
}
\label{fig:qmca_judge_opt}
\end{figure}
After inspecting the trace, we should inspect the text output
@ -274,15 +273,15 @@ cycles.
\begin{figure}
\centering
\begin{subfigure}[t]{0.47\textwidth}
\centering
\includegraphics[trim=0mm 0mm 4mm 0mm,clip,width=\linewidth]{figures/qmca_opt_energy.pdf}
\end{subfigure}
\begin{subfigure}[t]{0.47\textwidth}
\centering
\includegraphics[trim=2mm 0mm 4mm 0mm,clip,width=\linewidth]{figures/qmca_opt_variance.png}
\end{subfigure}
\caption{Energy and variance vs. optimization series for an 8 atom cell of diamond as plotted by \texttt{qmca}.\label{fig:qmca_opt_ev}}
\parbox{0.47\textwidth}{%
\includegraphics[trim=0mm 0mm 4mm 0mm,clip,width=\linewidth]{figures/qmca_opt_energy.pdf}%
}%
\qquad
\begin{minipage}{0.47\textwidth}%
\includegraphics[trim=2mm 0mm 4mm 0mm,clip,width=\linewidth]{figures/qmca_opt_variance.png}%
\end{minipage}%
\caption{Energy and variance vs. optimization series for an 8 atom cell of diamond as plotted by \texttt{qmca}.}%
\label{fig:qmca_opt_ev}%
\end{figure}
A good way to choose the optimal wavefunction for use in DMC is to select
@ -305,11 +304,10 @@ wavefunctions with larger than necessary variance.
\begin{figure}
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.75\columnwidth]{figures/qmca_short_dmc.pdf}
\includegraphics[trim = 0mm 0mm 0mm 0mm,clip,width=0.75\columnwidth]{figures/qmca_short_dmc.pdf}
\end{center}
\caption{Trace of the local energy for VMC followed by DMC with a small timestep ($0.002$ Ha$^{-1}$) for an 8 atom cell of diamond generated with \texttt{qmca}.
\caption{Trace of the local energy for VMC followed by DMC with a small timestep ($0.002$ Ha$^{-1}$) for an 8 atom cell of diamond generated with \texttt{qmca}.}
\label{fig:qmca_short_dmc}
}
\end{figure}
To illustrate the problems that can arise with respect to slow
@ -375,9 +373,8 @@ timestep--\emph{e.g.} $0.002$ Ha$^{-1}$--for a particular problem.
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.75\columnwidth]{figures/qmca_accel_dmc.pdf}
\end{center}
\caption{Trace of the local energy for VMC followed by a short intermediate DMC with a large timestep ($0.02$ Ha$^{-1}$) and finally a production DMC run with a timestep of $0.01$ Ha$^{-1}$. Calculations were performed in an 8 atom cell of diamond.
\caption{Trace of the local energy for VMC followed by a short intermediate DMC with a large timestep ($0.02$ Ha$^{-1}$) and finally a production DMC run with a timestep of $0.01$ Ha$^{-1}$. Calculations were performed in an 8 atom cell of diamond.}
\label{fig:qmca_accel_dmc}
}
\end{figure}
We now rerun the prior example but with an intermediate DMC
@ -429,9 +426,8 @@ Our final DMC total energy is estimated to be $-46.0329(2)$ Ha.
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.75\columnwidth]{figures/qmca_pop_trace.pdf}
\end{center}
\caption{Trace of the DMC walker population for an 8 atom cell of diamond obtained with \texttt{qmca}.
\caption{Trace of the DMC walker population for an 8 atom cell of diamond obtained with \texttt{qmca}.}
\label{fig:qmca_pop_trace}
}
\end{figure}
Another simulation property that should be explicitly monitored
@ -683,9 +679,8 @@ Fig. \ref{fig:qmca_twist_overlap}
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.9\columnwidth]{figures/qmca_twist_trace_overlap.pdf}
\end{center}
\caption{Overlapped energy traces from VMC to DMC for an 8 supercell of diamond obtained with \texttt{qmca}. Data for each twist appears in a different color.
\caption{Overlapped energy traces from VMC to DMC for an 8 supercell of diamond obtained with \texttt{qmca}. Data for each twist appears in a different color.}
\label{fig:qmca_twist_overlap}
}
\end{figure}
Twist averaging is performed by providing the ``\texttt{-a}''
@ -719,9 +714,8 @@ yielding a twist averaged total energy of $-45.8733(8)$ Ha.
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.75\columnwidth]{figures/qmca_twist_average_trace.pdf}
\end{center}
\caption{Twist averaged energy trace from VMC to DMC for an 8 supercell of diamond obtained with \texttt{qmca}.
\caption{Twist averaged energy trace from VMC to DMC for an 8 supercell of diamond obtained with \texttt{qmca}.}
\label{fig:qmca_twist_average}
}
\end{figure}
As can be seen from the Fig. \ref{fig:qmca_twist_overlap}, some of the twist
@ -1120,15 +1114,15 @@ panel of Fig.~\ref{fig:qfit_timestep}.
\begin{figure}
\centering
\begin{subfigure}[t]{0.47\textwidth}
\centering
\includegraphics[trim=0mm 0mm 4mm 0mm,clip,width=\linewidth]{figures/qfit_timestep_linear.pdf}
\end{subfigure}
\begin{subfigure}[t]{0.47\textwidth}
\centering
\includegraphics[trim=2mm 0mm 4mm 0mm,clip,width=\linewidth]{figures/qfit_timestep_quadratic.pdf}
\end{subfigure}
\caption{Linear (left) and quadratic (right) timestep fits to DMC data for a 32 atom supercell of MnO obtained with \texttt{qfit}. Zero timestep estimates are indicated by the red data point on the left side of either panel.\label{fig:qfit_timestep}}
\parbox{0.47\textwidth}{%
\includegraphics[trim=0mm 0mm 4mm 0mm,clip,width=\linewidth]{figures/qfit_timestep_linear.pdf}%
}%
\qquad
\begin{minipage}{0.47\textwidth}%
\includegraphics[trim=2mm 0mm 4mm 0mm,clip,width=\linewidth]{figures/qfit_timestep_quadratic.pdf}%
\end{minipage}%
\caption{Linear (left) and quadratic (right) timestep fits to DMC data for a 32 atom supercell of MnO obtained with \texttt{qfit}. Zero timestep estimates are indicated by the red data point on the left side of either panel.}
\label{fig:qfit_timestep}
\end{figure}
Different fitting functions are supported via the ``\texttt{-f}'' option.
@ -1161,3 +1155,5 @@ estimated value of $-3848.28(7)$ instead.
\label{sec:energydensities}

3
manual/design_features.tex
View File

@ -609,7 +609,8 @@ S =
We note at the left and right extremes, the values and first two
derivatives of the functions are zero, while at the center, $h_{j0}$
has a value of 1, $h_{j1}$ has a first derivative of 1, and $h_{j2}$
has a second derivative of 1. \label{fig:LPQHI} }
has a second derivative of 1. }
\label{fig:LPQHI}
\end{center}
\end{figure}
Figure~\ref{fig:LPQHI} shows plots of these function shapes.

3
manual/hamiltonianobservable.tex
View File

@ -989,7 +989,8 @@ Additional information
\hline
\end{tabular}
\end{center}
\caption{Reference points available by default. The vectors $a_1$, $a_2$, and $a_3$ refer to the simulation cell axes. The representation of the cell is centered around \texttt{zero}. \label{tab:ref_points}}
\caption{Reference points available by default. The vectors $a_1$, $a_2$, and $a_3$ refer to the simulation cell axes. The representation of the cell is centered around \texttt{zero}.}
\protect{\label{tab:ref_points}}
\end{table}
\FloatBarrier

3
manual/installation.tex
View File

@ -1311,8 +1311,7 @@ the following combinations nightly (workstations) and weekly (supercomputers):
\caption{Example test results for QMCPACK, showing data for a
workstation (Intel, GCC, both CPU and GPU builds) and for two ORNL
supercomputers. In this example, 4 errors were found. This
dashboard is accessible at \url{https://cdash.qmcpack.org/}.}
\label{fig:cdash}
dashboard is accessible at \url{https://cdash.qmcpack.org}.} %%\label{fig:cdash}}
\end{figure}
\section{Building ppconvert, a pseudopotential format converter}

36
manual/lab_advanced_molecules.tex
View File

@ -167,9 +167,8 @@ qmca -q e *scalar.dat -p
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.75\columnwidth]{figures/lab_advanced_molecules_opt_conv.png}
\end{center}
\caption{VMC energy as a function of optimization step.
\caption{VMC energy as a function of optimization step.}
\label{fig:lam_opt_conv}
}
\end{figure}
The resulting energy as a function of optimization step should look qualitatively similar to figure \ref{fig:lam_opt_conv}.
@ -219,9 +218,8 @@ limit.
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.75\columnwidth]{figures/lab_advanced_molecules_dmc_timestep.png}
\end{center}
\caption{DMC energy as a function of timestep.
\caption{DMC energy as a function of timestep.}
\label{fig:lam_dmc_timestep}
}
\end{figure}
@ -267,9 +265,8 @@ jobrun_vesta qmcpack dmc_wk.xml
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.75\columnwidth]{figures/lab_advanced_molecules_dmc_popcont.png}
\end{center}
\caption{DMC energy as a function of the average number of walkers.
\caption{DMC energy as a function of the average number of walkers.}
\label{fig:lam_dmc_popcont}
}
\end{figure}
Repeat the same procedure in the other folders by setting (targetWalkers=240,
@ -333,9 +330,8 @@ input, just leave the VMC step. In all three cases, modify the submission script
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.75\columnwidth]{figures/lab_advanced_molecules_vmc_jastrow.png}
\end{center}
\caption{VMC energy as a function of Jastrow type.
\caption{VMC energy as a function of Jastrow type.}
\label{fig:lam_vmc_jastrow}
}
\end{figure}
These simulations will take several minutes to complete. This is an excellent opportunity
@ -437,9 +433,8 @@ After the wave-function is generated, we are ready to optimize. Instead of start
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.75\columnwidth]{figures/lab_advanced_molecules_dmc_ci_cisd.png}
\end{center}
\caption{DMC energy as a function of the sum of the square of CI coefficients from CISD.
\caption{DMC energy as a function of the sum of the square of CI coefficients from CISD.}
\label{fig:lam_dmc_ci_cisd}
}
\end{figure}
When you are done, use qmca to analyze the results. Compare the energies at these two
@ -746,9 +741,8 @@ symmetry rotation, the resulting output will always be consistent.}
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.70\columnwidth]{figures/lab_advanced_molecules_xml_opt.png}
\end{center}
\caption{Sample XML optimization block.
\caption{Sample XML optimization block.}
\label{fig:lam_xml_opt}
}
\end{figure}
%\FloatBarrier
@ -794,9 +788,8 @@ to 0 and do a single inner iteration (max its=1) per sample of configurations.}
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.75\columnwidth]{figures/lab_advanced_molecules_xml_vmc_dmc.png}
\end{center}
\caption{Sample XML blocks for VMC and DMC calculations.
\caption{Sample XML blocks for VMC and DMC calculations.}
\label{fig:lam_xml_vmc_dmc}
}
\end{figure}
%\FloatBarrier
@ -841,9 +834,8 @@ given node, the walkers are split across all the threads in the system.}
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=1.0\columnwidth]{figures/lab_advanced_molecules_xml_determinantset.png}
\end{center}
\caption{Basic framework for a single determinant determinantset XML block.
\caption{Basic framework for a single determinant determinantset XML block.}
\label{fig:lam_xml_determinantset}
}
\end{figure}
%\FloatBarrier
@ -859,9 +851,8 @@ functions) since changes can lead to unexpected results.
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=1.0\columnwidth]{figures/lab_advanced_molecules_xml_basisset.png}
\end{center}
\caption{Sample XML block for an atomic orbital basis set.
\caption{Sample XML block for an atomic orbital basis set.}
\label{fig:lam_xml_basisset}
}
\end{figure}
%\FloatBarrier
@ -887,9 +878,8 @@ basis functions per single particle orbital.
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=1.0\columnwidth]{figures/lab_advanced_molecules_xml_slaterdeterminant.png}
\end{center}
\caption{Sample XML block for the single slater determinant case.
\caption{Sample XML block for the single slater determinant case.}
\label{fig:lam_xml_slaterdeterminant}
}
\end{figure}
%\FloatBarrier
@ -908,9 +898,8 @@ terms of occupation numbers based on these orbitals.
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=1.0\columnwidth]{figures/lab_advanced_molecules_xml_multideterminant.png}
\end{center}
\caption{Basic framework for a multi-determinant determinantset XML block.
\caption{Basic framework for a multi-determinant determinantset XML block.}
\label{fig:lam_xml_multideterminant}
}
\end{figure}
%\FloatBarrier
@ -944,9 +933,8 @@ present in the J1 and J2 blocks.
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=1.0\columnwidth]{figures/lab_advanced_molecules_xml_jastrow.png}
\end{center}
\caption{Sample Jastrow XML block.
\caption{Sample Jastrow XML block.}
\label{fig:lam_xml_jastrow}
}
\end{figure}
%\FloatBarrier

3
manual/lab_qmc_basics.tex
View File

@ -619,9 +619,8 @@ b = -0.0457479 +/- 0.0422 (92.24%)
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.75\columnwidth]{figures/lab_qmc_basics_timestep_conv.pdf}
\end{center}
\caption{Linear fit to DMC timestep data from \texttt{PlotTstepConv.pl}.
\caption{Linear fit to DMC timestep data from \texttt{PlotTstepConv.pl}.}
\label{fig:timestep_conv}
}
\end{figure}

2
manual/opt.tex
View File

@ -369,7 +369,7 @@ This optimizer is very robust but a bit conservative to accept new steps especia
Additional information:
\begin{itemize}
\item \texttt{exp0}. It is the initial value for stabilizer (shift to diagonal of H). The actual value of stabilizer is $10^\textrm{exp0}$.
\item \texttt{exp0}. It is the initial value for stabilizer (shift to diagonal of H). The actual value of stabilizer is $10^{\textrm{exp0}}$.
\end{itemize}
Recommendations:

6
manual/qmcpack_manual.tex
View File

@ -15,7 +15,7 @@
% environment for shaded verbatim
\usepackage{verbatim}
\usepackage{caption}
\usepackage{subcaption}
%\usepackage{subcaption}
\usepackage{graphicx}
\usepackage{epsfig}
% making listing behave properly
@ -213,14 +213,14 @@
\input{design_features}
\input{developing}
\chapter{References}
%\chapter{References}
%\begingroup % Avoid "References" section title
%\renewcommand{\section}[2]{}%
%\bibliographystyle{ieeetr}
%\bibliography{bibliography}
%\endgroup
\btPrintCited
%%\btPrintCited
\end{btSect}
\end{btUnit}

16
manual/selected_ci.tex
View File

@ -106,9 +106,8 @@ It is important to note that:
\begin{center}
\includegraphics[trim = 0mm 0mm 0mm 0mm, clip,width=0.3\columnwidth]{figures/Reactant.jpg}
\end{center}
\caption{$C_2O_2H_3N$ molecule.
\label{fig:C2O2H3N}
}
\caption{$C_2O_2H_3N$ molecule.}
\protect{\label{fig:C2O2H3N}}
\end{figure}
The CIPSI method
%\cite{XXXrecentCIPSIpaper}
@ -204,9 +203,8 @@ At this point, the orbitals are modified, a new AO$\rightarrow$MO transformation
\begin{center}
\includegraphics[trim = 2mm 2mm 2mm 2mm, clip,width=0.95\columnwidth]{figures/CIPSI.jpg}
\end{center}
\caption{Evolution of the variational energy and the Energy + PT2 as a function of the number of determinants for the $C_2O_2H_3N$ molecule.
\label{fig:CIPSI}
}
\caption{Evolution of the variational energy and the Energy + PT2 as a function of the number of determinants for the $C_2O_2H_3N$ molecule.}
\protect{\label{fig:CIPSI}}
\end{figure}
Figure \ref{fig:CIPSI} shows the evolution of the variational energy and the energy corrected with PT2 as a function of the number of determinants up to 4M determinants. While it is clear that the raw variational energy is far from being converged, the Energy + PT2 appears converged around 0.4M determinants.\\
@ -355,7 +353,7 @@ ideal benchmark tool for multireference systems.
\includegraphics[trim = 2mm 2mm 2mm 2mm, clip,width=0.9
\columnwidth]{figures/DMC-Multidet.jpg}
\end{center}
\caption{DMC energy of $C_2O_2H_3N$ molecule as a function of different single determinant trial wavefunctions with aug-ccp-VTZ basis set using nodal surfaces from Hartree-Fock (HF), DFT-PBE and DFT with hybrid functionals PBE0 and P3LYP. The CIPSI trial wavefunction contains as indicated 239, 44539, 514380 and 908128 determinants (D).
\label{fig:CIPSI-DMC}
}
\caption{DMC energy of $C_2O_2H_3N$ molecule as a function of different single determinant trial wavefunctions with aug-ccp-VTZ basis set using nodal surfaces from Hartree-Fock (HF), DFT-PBE and DFT with hybrid functionals PBE0 and P3LYP. The CIPSI trial wavefunction contains as indicated 239, 44539, 514380 and 908128 determinants (D). }
\protect{\label{fig:CIPSI-DMC}}
\end{figure}