\par
\section{Driver programs for the {\tt Chv object}}
\label{section:Chv:drivers}
\par
%=======================================================================
\begin{enumerate}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
test_addChevron msglvl msgFile nD nU type symflag seed alphareal alphaimag
\end{verbatim}
This driver program tests the {\tt Chv\_addChevron} method.
Use the script file {\tt do\_addChevron} for testing.
When the output file is loaded into matlab,
the last line to the screen is the error of the assembly.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
Use {\tt msglvl = 1} for just timing output.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt nD} parameter is the number of rows and columns 
in the (1,1) block.
\item
The {\tt nU} parameter is the number of columns 
in the (1,2) block.
\item
The {\tt type} parameter denotes the type of entries ---
{\tt SPOOLES\_REAL} or {\tt SPOOLES\_COMPLEX} 
\item
The {\tt symflag} parameter is the symmetry flag ---
{\tt SPOOLES\_SYMMETRIC}, {\tt SPOOLES\_HERMITIAN}
or {\tt SPOOLES\_NONSYMMETRIC}.
\item
The {\tt seed} parameter is a random number seed.
\item
The {\tt alphareal} and {\tt alphaimag} parameters form a complex
number that is a scaling parameter.
Normally {\tt alpha} is (1.0,0.0), when we are just loading matrix
entries into a front.
However, when we factor $A + \alpha B$, the entries of $B$ will
be loaded with {\tt alpha} set equal to $\alpha[0:1]$.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
test_assmbChv msglvl msgFile nDJ nUJ nDI nUI type symflag seed
\end{verbatim}
This driver program tests the {\tt Chv\_assembleChv} method.
It assembles a chevron $T_I$ into $T_J$, as is done during the
assembly of postponed rows and columns during the factorization
when pivoting is enabled.
Use the script file {\tt do\_assmbChv} for testing.
When the output file is loaded into matlab,
the last line to the screen is the error of the assembly.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
Use {\tt msglvl = 1} for just timing output.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt nDJ} parameter is the number of rows and columns 
in the (1,1) block of $T_J$.
\item
The {\tt nUJ} parameter is the number of columns 
in the (1,2) block of $T_J$.
\item
The {\tt nDI} parameter is the number of rows and columns 
in the (1,1) block of $T_I$.
\item
The {\tt nUI} parameter is the number of columns 
in the (1,2) block of $T_I$.
\item
The {\tt type} parameter denotes the type of entries ---
{\tt SPOOLES\_REAL} or {\tt SPOOLES\_COMPLEX} 
\item
The {\tt symflag} parameter is the symmetry flag ---
{\tt SPOOLES\_SYMMETRIC}, {\tt SPOOLES\_HERMITIAN}
or {\tt SPOOLES\_NONSYMMETRIC}.
\item
The {\tt seed} parameter is a random number seed.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
test_copyEntriesToVector msglvl msgFile nD nU type symflag 
                         pivotingflag storeflag seed 
\end{verbatim}
This driver program tests the {\tt Chv\_copyEntriesToVector} method
which is used when after a front has been factored to store the
entries into dense $L$ and $U$ submatrices.
Use the script file {\tt do\_copyEntriesToVector} for testing.
When the output file is loaded into matlab,
the last line to the screen is a matrix that contains two entries.
If the program executes correctly, these two entries should be
exactly zero.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
Use {\tt msglvl = 1} for just timing output.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt nD} parameter is the number of rows and columns 
in the (1,1) block.
\item
The {\tt nU} parameter is the number of columns 
in the (1,2) block.
\item
The {\tt type} parameter denotes the type of entries ---
{\tt SPOOLES\_REAL} or {\tt SPOOLES\_COMPLEX} 
\item
The {\tt symflag} parameter is the symmetry flag ---
{\tt SPOOLES\_SYMMETRIC}, {\tt SPOOLES\_HERMITIAN}
or {\tt SPOOLES\_NONSYMMETRIC}.
\item
The {\tt pivotingflag} parameter is the pivoting flag ---
{\tt SPOOLES\_NO\_PIVOTING} for no pivoting,
{\tt SPOOLES\_PIVOTING} for pivoting.
\item
The {\tt storeflag} parameter is the storage flag,
to store by rows, use {\tt SPOOLES\_BY\_ROWS},
to store by columns, use {\tt SPOOLES\_BY\_COLUMNS}.
\item
The {\tt seed} parameter is a random number seed.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
test_copyBigEntriesToVector msglvl msgFile nD nU type symflag 
                         pivotingflag storeflag seed droptol
\end{verbatim}
This driver program tests the {\tt Chv\_copyBigEntriesToVector} method
which is used when after a front has been factored to store the
entries into sparse $L$ and $U$ submatrices.
Use the script file {\tt do\_copyBigEntriesToVector} for testing.
When the output file is loaded into matlab,
the last line to the screen is a matrix that contains three entries.
The first two are the maximum magnitudes of the entries that were
not copied (two different ways), and the third is the drop tolerance.
If the program executes correctly, the third term is larger than
the first two.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
Use {\tt msglvl = 1} for just timing output.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt nD} parameter is the number of rows and columns 
in the (1,1) block.
\item
The {\tt nU} parameter is the number of columns 
in the (1,2) block.
\item
The {\tt type} parameter denotes the type of entries ---
{\tt SPOOLES\_REAL} or {\tt SPOOLES\_COMPLEX} 
\item
The {\tt symflag} parameter is the symmetry flag ---
{\tt SPOOLES\_SYMMETRIC}, {\tt SPOOLES\_HERMITIAN}
or {\tt SPOOLES\_NONSYMMETRIC}.
\item
The {\tt pivotingflag} parameter is the pivoting flag ---
{\tt SPOOLES\_NO\_PIVOTING} for no pivoting,
{\tt SPOOLES\_PIVOTING} for pivoting.
\item
The {\tt storeflag} parameter is the storage flag,
to store by rows, use {\tt SPOOLES\_BY\_ROWS},
to store by columns, use {\tt SPOOLES\_BY\_COLUMNS}.
\item
The {\tt seed} parameter is a random number seed.
\item
The {\tt droptol} parameter is a drop tolerance parameters,
entries whose magnitude is smaller than {\tt droptol} are not
copied.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
test_factor msglvl msgFile nD nU type symflag pivotingflag seed tau
\end{verbatim}
This driver program tests the {\tt Chv\_factor} method.
Use the script file {\tt do\_factor} for testing.
When the output file is loaded into matlab,
the last line to the screen is a matrix that contains three entries.
The first entry is the error in the factorization.
The second and third entries are the maximum magnitudes of the
entries in $L$ and $U$, respectively.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
Use {\tt msglvl = 1} for just timing output.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt nD} parameter is the number of rows and columns 
in the (1,1) block.
\item
The {\tt nU} parameter is the number of columns 
in the (1,2) block.
\item
The {\tt type} parameter denotes the type of entries ---
{\tt SPOOLES\_REAL} or {\tt SPOOLES\_COMPLEX} 
\item
The {\tt symflag} parameter is the symmetry flag ---
{\tt SPOOLES\_SYMMETRIC}, {\tt SPOOLES\_HERMITIAN}
or {\tt SPOOLES\_NONSYMMETRIC}.
\item
The {\tt pivotingflag} parameter is the pivoting flag ---
{\tt SPOOLES\_NO\_PIVOTING} for no pivoting,
{\tt SPOOLES\_PIVOTING} for pivoting.
\item
The {\tt seed} parameter is a random number seed.
\item
The {\tt tau} parameter is used when pivoting is enabled.
All entries in $L$ and $U$ will have magnitudes less than {\tt tau}.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
test_findPivot msglvl msgFile nD nU type symflag seed tau
\end{verbatim}
This driver program tests the {\tt Chv\_findPivot} method.
Use the script file {\tt do\_findPivot} for testing.
When the output file is loaded into matlab,
look on the screen for the variables
{\tt maxerrupd} (the error in the factor and update),
{\tt ubound} (the maximum magnitude of the entries in $U$),
and if nonsymmetric
{\tt lbound} (the maximum magnitude of the entries in $L$).
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
Use {\tt msglvl = 1} for just timing output.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt nD} parameter is the number of rows and columns 
in the (1,1) block.
\item
The {\tt nU} parameter is the number of columns 
in the (1,2) block.
\item
The {\tt type} parameter denotes the type of entries ---
{\tt SPOOLES\_REAL} or {\tt SPOOLES\_COMPLEX} 
\item
The {\tt symflag} parameter is the symmetry flag ---
{\tt SPOOLES\_SYMMETRIC}, {\tt SPOOLES\_HERMITIAN}
or {\tt SPOOLES\_NONSYMMETRIC}.
\item
The {\tt seed} parameter is a random number seed.
\item
The {\tt tau} parameter is used when pivoting is enabled.
All entries in $L$ and $U$ will have magnitudes less than {\tt tau}.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
test_maxabs msglvl msgFile nD nU type symflag seed 
\end{verbatim}
This driver program tests the 
{\tt Chv\_maxabsInRow()},
{\tt Chv\_maxabsInRow11()},
{\tt Chv\_maxabsInColumn()},
{\tt Chv\_maxabsInColumn11()} and
{\tt Chv\_maxabsInDiagonal11()} 
methods.
Use the script file {\tt do\_maxabs} for testing.
When the output file is loaded into matlab,
look on the screen for the variables
{\tt rowerror}, {\tt colerror},
{\tt rowerror11}, {\tt colerror11}
and {\tt diag11error}.
All should be zero.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
Use {\tt msglvl = 1} for just timing output.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt nD} parameter is the number of rows and columns 
in the (1,1) block.
\item
The {\tt nU} parameter is the number of columns 
in the (1,2) block.
\item
The {\tt type} parameter denotes the type of entries ---
{\tt SPOOLES\_REAL} or {\tt SPOOLES\_COMPLEX} 
\item
The {\tt symflag} parameter is the symmetry flag ---
{\tt SPOOLES\_SYMMETRIC}, {\tt SPOOLES\_HERMITIAN}
or {\tt SPOOLES\_NONSYMMETRIC}.
\item
The {\tt pivotingflag} parameter is the pivoting flag ---
{\tt SPOOLES\_NO\_PIVOTING} for no pivoting,
{\tt SPOOLES\_PIVOTING} for pivoting.
\item
The {\tt seed} parameter is a random number seed.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
test_r1upd msglvl msgFile nD nU type symflag seed 
\end{verbatim}
This driver program tests the {\tt Chv\_r1upd()} method.
Use the script file {\tt do\_r1upd} for testing.
When the output file is loaded into matlab,
the last line is the error of the update.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
Use {\tt msglvl = 1} for just timing output.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt nD} parameter is the number of rows and columns 
in the (1,1) block.
\item
The {\tt nU} parameter is the number of columns 
in the (1,2) block.
\item
The {\tt type} parameter denotes the type of entries ---
{\tt SPOOLES\_REAL} or {\tt SPOOLES\_COMPLEX} 
\item
The {\tt symflag} parameter is the symmetry flag ---
{\tt SPOOLES\_SYMMETRIC}, {\tt SPOOLES\_HERMITIAN}
or {\tt SPOOLES\_NONSYMMETRIC}.
\item
The {\tt seed} parameter is a random number seed.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
test_r2upd msglvl msgFile nD nU type symflag seed 
\end{verbatim}
This driver program tests the {\tt Chv\_r2upd()} method.
Use the script file {\tt do\_r2upd} for testing.
When the output file is loaded into matlab,
the last line is the error of the update.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
Use {\tt msglvl = 1} for just timing output.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt nD} parameter is the number of rows and columns 
in the (1,1) block.
\item
The {\tt nU} parameter is the number of columns 
in the (1,2) block.
\item
The {\tt type} parameter denotes the type of entries ---
{\tt SPOOLES\_REAL} or {\tt SPOOLES\_COMPLEX} 
\item
The {\tt symflag} parameter is the symmetry flag ---
{\tt SPOOLES\_SYMMETRIC}, {\tt SPOOLES\_HERMITIAN}
or {\tt SPOOLES\_NONSYMMETRIC}.
\item
The {\tt seed} parameter is a random number seed.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
test_swap msglvl msgFile nD nU type symflag seed 
\end{verbatim}
This driver program tests three methods: 
{\tt Chv\_swapRowsAndColumns()},
{\tt Chv\_swapRows()} and
{\tt Chv\_swapColumns()}.
Use the script file {\tt do\_swap} for testing.
When the output file is loaded into matlab,
look for the {\tt maxerrrowswap1}, {\tt maxerrcolswap1},
{\tt maxerrswap}, {\tt maxerrsymswap1} and {\tt maxerrsymswap2} values.
All should be zero.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
Use {\tt msglvl = 1} for just timing output.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt nD} parameter is the number of rows and columns 
in the (1,1) block.
\item
The {\tt nU} parameter is the number of columns 
in the (1,2) block.
\item
The {\tt type} parameter denotes the type of entries ---
{\tt SPOOLES\_REAL} or {\tt SPOOLES\_COMPLEX} 
\item
The {\tt symflag} parameter is the symmetry flag ---
{\tt SPOOLES\_SYMMETRIC}, {\tt SPOOLES\_HERMITIAN}
or {\tt SPOOLES\_NONSYMMETRIC}.
\item
The {\tt seed} parameter is a random number seed.
\end{itemize}
%-----------------------------------------------------------------------
\item
\begin{verbatim}
test_update msglvl msgFile type symflag sparsityflag
            ncolT ncolU nrowD nentU offset seed 
\end{verbatim}
This driver program tests the 
{\tt Chv\_updateH()},
{\tt Chv\_updateS()} and
{\tt Chv\_updateN()} methods.
The {\tt Chv} object $T$ is updated by $-U^TDU$, $-U^HDU$ or $-LDU$, 
depending on whether $T$ is symmetric, hermitian or nonsymmetric.
Use the script file {\tt do\_update} for testing.
When the output file is loaded into matlab,
the last line is the error in the update which should be zero.
\par
\begin{itemize}
\item
The {\tt msglvl} parameter determines the amount of output.
Use {\tt msglvl = 1} for just timing output.
\item
The {\tt msgFile} parameter determines the message file --- if {\tt
msgFile} is {\tt stdout}, then the message file is {\it stdout},
otherwise a file is opened with {\it append} status to receive any
output data.
\item
The {\tt type} parameter denotes the type of entries ---
{\tt SPOOLES\_REAL} or {\tt SPOOLES\_COMPLEX} 
\item
The {\tt symflag} parameter is the symmetry flag ---
{\tt SPOOLES\_SYMMETRIC}, {\tt SPOOLES\_HERMITIAN}
or {\tt SPOOLES\_NONSYMMETRIC}.
\item
The {\tt sparsityflag} parameter should be zero for dense $U$ and $L$,
or 1 for sparse $U$ and $L$.
\item
The {\tt ncolT} parameter is the number of columns 
in the (1,1) and (1,2) blocks of $T$.
\item
The {\tt nDT} parameter is the number of rows and columns 
in the (1,1) block of $T$.
\item
The {\tt ncolU} parameter is the number of columns in $U$.
\item
The {\tt nrowD} parameter is the number of rows and columns in $D$.
\item
The {\tt nentU} parameter is the number entries in $U$, ignored if
{\tt sparsityflag = 0}.
\item
The {\tt offset} parameter is the offset of first index in $T$ from
the last index in $D$.
\item
The {\tt seed} parameter is a random number seed.
\end{itemize}
%-----------------------------------------------------------------------
\end{enumerate}