This is really old school, but a lot of times we have users sending us
copy pasting of codes, and that is the only way to know the version of
the code.
Close#82
Added routines for new factorization code for symmetric indefinite
( or Hermitian indefinite ) matrices with bounded Bunch-Kaufman
( rook ) pivoting algorithm.
New more efficient storage format for factors U ( or L ),
block-diagonal matrix D, and pivoting information stored in IPIV:
factor L is stored explicitly in lower triangle of A;
diagonal of D is stored on the diagonal of A;
subdiagonal elements of D are stored in array E;
IPIV format is the same as in *_ROOK routines, but differs
from SY Bunch-Kaufman routines (e.g. *SYTRF).
The factorization output of these new rook _RK routines is not
compatible
with the existing _ROOK routines and vice versa. This new factorization
format is designed in such a way, that there is a possibility in the
future
to write new Bunch-Kaufman routines that conform to this new
factorization format.
Then the future Bunch-Kaufman routines could share solver
*TRS_3,inversion *TRI_3
and condition estimator *CON_3.
To convert between the factorization formats in both ways the following
routines
are developed:
CONVERSION ROUTINES BETWEEN FACTORIZATION FORMATS
DOUBLE PRECISION (symmetric indefinite matrices):
new file: SRC/dsyconvf.f
new file: SRC/dsyconvf_rook.f
REAL (symmetric indefinite matrices):
new file: SRC/csyconvf.f
new file: SRC/csyconvf_rook.f
COMPLEX*16 (symmetric indefinite and Hermitian indefinite matrices):
new file: SRC/zsyconvf.f
new file: SRC/zsyconvf_rook.f
COMPLEX (symmetric indefinite and Hermitian indefinite matrices):
new file: SRC/ssyconvf.f
new file: SRC/ssyconvf_rook.f
*SYCONVF routine converts between old Bunch-Kaufman storage format (
denote (L1,D1,IPIV1) )
that is used by *SYTRF and new rook storage format ( denote (L2,D2,
IPIV2))
that is used by *SYTRF_RK
*SYCONVF_ROOK routine between old rook storage format ( denote
(L1,D1,IPIV2) )
that is used by *SYTRF_ROOK and new rook storage format ( denote
(L2,D2, IPIV2))
that is used by *SYTRF_RK
ROUTINES AND DRIVERS
DOUBLE PRECISION (symmetric indefinite matrices):
new file: SRC/dsytf2_rk.f BLAS2 unblocked factorization
new file: SRC/dlasyf_rk.f BLAS3 auxiliary blocked partial
factorization
new file: SRC/dsytrf_rk.f BLAS3 blocked factorization
new file: SRC/dsytrs_3.f BLAS3 solver
new file: SRC/dsycon_3.f BLAS3 condition number estimator
new file: SRC/dsytri_3.f BLAS3 inversion, sets the size of work array
and calls *sytri_3x
new file: SRC/dsytri_3x.f BLAS3 auxiliary inversion, actually
computes blocked inversion
new file: SRC/dsysv_rk.f BLAS3 solver driver
REAL (symmetric indefinite matrices):
new file: SRC/ssytf2_rk.f BLAS2 unblocked factorization
new file: SRC/slasyf_rk.f BLAS3 auxiliary blocked partial
factorization
new file: SRC/ssytrf_rk.f BLAS3 blocked factorization
new file: SRC/ssytrs_3.f BLAS3 solver
new file: SRC/ssycon_3.f BLAS3 condition number estimator
new file: SRC/ssytri_3.f BLAS3 inversion, sets the size of work array
and calls *sytri_3x
new file: SRC/ssytri_3x.f BLAS3 auxiliary inversion, actually
computes blocked inversion
new file: SRC/ssysv_rk.f BLAS3 solver driver
COMPLEX*16 (symmetric indefinite matrices):
new file: SRC/zsytf2_rk.f BLAS2 unblocked factorization
new file: SRC/zlasyf_rk.f BLAS3 auxiliary blocked partial
factorization
new file: SRC/zsytrf_rk.f BLAS3 blocked factorization
new file: SRC/zsytrs_3.f BLAS3 solver
new file: SRC/zsycon_3.f BLAS3 condition number estimator
new file: SRC/zsytri_3.f BLAS3 inversion, sets the size of work array
and calls *sytri_3x
new file: SRC/zsytri_3x.f BLAS3 auxiliary inversion, actually
computes blocked inversion
new file: SRC/zsysv_rk.f BLAS3 solver driver
COMPLEX*16 (Hermitian indefinite matrices):
new file: SRC/zhetf2_rk.f BLAS2 unblocked factorization
new file: SRC/zlahef_rk.f BLAS3 auxiliary blocked partial
factorization
new file: SRC/zhetrf_rk.f BLAS3 blocked factorization
new file: SRC/zhetrs_3.f BLAS3 solver
new file: SRC/zhecon_3.f BLAS3 condition number estimator
new file: SRC/zhetri_3.f BLAS3 inversion, sets the size of work array
and calls *sytri_3x
new file: SRC/zhetri_3x.f BLAS3 auxiliary inversion, actually
computes blocked inversion
new file: SRC/zhesv_rk.f BLAS3 solver driver
COMPLEX (symmetric indefinite matrices):
new file: SRC/csytf2_rk.f BLAS2 unblocked factorization
new file: SRC/clasyf_rk.f BLAS3 auxiliary blocked partial
factorization
new file: SRC/csytrf_rk.f BLAS3 blocked factorization
new file: SRC/csytrs_3.f BLAS3 solver
new file: SRC/csycon_3.f BLAS3 condition number estimator
new file: SRC/csytri_3.f BLAS3 inversion, sets the size of work array
and calls *sytri_3x
new file: SRC/csytri_3x.f BLAS3 auxiliary inversion, actually
computes blocked inversion
new file: SRC/csysv_rk.f BLAS3 solver driver
COMPLEX (Hermitian indefinite matrices):
new file: SRC/chetf2_rk.f BLAS2 unblocked factorization
new file: SRC/clahef_rk.f BLAS3 auxiliary blocked partial
factorization
new file: SRC/chetrf_rk.f BLAS3 blocked factorization
new file: SRC/chetrs_3.f BLAS3 solver
new file: SRC/checon_3.f BLAS3 condition number estimator
new file: SRC/chetri_3.f BLAS3 inversion, sets the size of work array
and calls *sytri_3x
new file: SRC/chetri_3x.f BLAS3 auxiliary inversion, actually
computes blocked inversion
new file: SRC/chesv_rk.f BLAS3 solver driver
MISC
modified: SRC/CMakeLists.txt
modified: SRC/Makefile
TEST CODE
modified: TESTING/LIN/CMakeLists.txt
modified: TESTING/LIN/Makefile
modified: TESTING/LIN/aladhd.f
modified: TESTING/LIN/alaerh.f
modified: TESTING/LIN/alahd.f
DOUBLE PRECISION (symmetric indefinite matrices):
modified: TESTING/LIN/dchkaa.f
modified: TESTING/LIN/derrsy.f
modified: TESTING/LIN/derrsyx.f
modified: TESTING/LIN/derrvx.f
modified: TESTING/LIN/derrvxx.f
modified: TESTING/dtest.in
new file: TESTING/LIN/dchksy_rk.f
new file: TESTING/LIN/ddrvsy_rk.f
new file: TESTING/LIN/dsyt01_3.f
REAL (symmetric indefinite matrices):
modified: TESTING/LIN/schkaa.f
modified: TESTING/LIN/serrsy.f
modified: TESTING/LIN/serrsyx.f
modified: TESTING/LIN/serrvx.f
modified: TESTING/LIN/serrvxx.f
modified: TESTING/stest.in
new file: TESTING/LIN/schksy_rk.f
new file: TESTING/LIN/sdrvsy_rk.f
new file: TESTING/LIN/ssyt01_3.f
COMPLEX*16 (symmetric indefinite and Hermitian indefinite matrices):
modified: TESTING/LIN/zchkaa.f
modified: TESTING/LIN/zerrsy.f
modified: TESTING/LIN/zerrsyx.f
modified: TESTING/LIN/zerrhe.f
modified: TESTING/LIN/zerrhex.f
modified: TESTING/LIN/zerrvx.f
modified: TESTING/LIN/zerrvxx.f
modified: TESTING/ztest.in
new file: TESTING/LIN/zchksy_rk.f
new file: TESTING/LIN/zdrvsy_rk.f
new file: TESTING/LIN/zsyt01_3.f
new file: TESTING/LIN/zchkhe_rk.f
new file: TESTING/LIN/zdrvhe_rk.f
new file: TESTING/LIN/zhet01_3.f
COMPLEX (symmetric indefinite and Hermitian indefinite matrices):
modified: TESTING/LIN/cchkaa.f
modified: TESTING/LIN/cerrsy.f
modified: TESTING/LIN/cerrsyx.f
modified: TESTING/LIN/cerrhe.f
modified: TESTING/LIN/cerrhex.f
modified: TESTING/LIN/cerrvx.f
modified: TESTING/LIN/cerrvxx.f
modified: TESTING/ctest.in
new file: TESTING/LIN/cchksy_rk.f
new file: TESTING/LIN/cdrvsy_rk.f
new file: TESTING/LIN/csyt01_3.f
new file: TESTING/LIN/cchkhe_rk.f
new file: TESTING/LIN/cdrvhe_rk.f
new file: TESTING/LIN/chet01_3.f
This is mostly a long term maintenance improvement.
Many coding styles require elimination of trailing whitespace, and
many editors and source code management configurations automatically
gobble up whitespace. When these tools gobble up whitespace, it
complicates reviewing the meaningful code changes.
By removing whitespace on one patch, it makes future
code reviews much easier.
=SCRIPT====================================================================
if which tempfile &>/dev/null; then
TEMPMAKER=tempfile
elif which mktemp &>/dev/null; then
TEMPMAKER=mktemp
else
echo "Cannot find tempfile program." 2>&1
exit 1
fi
MYTEMP=$($TEMPMAKER)
trap 'rm -f $MYTEMP' SIGINT SIGTERM
stripit() {
echo "stripping $1"
sed 's/[ \t]*$//' "$1" > $MYTEMP
cp $MYTEMP "$1"
}
if [ $# -gt 0 ]; then
while [ "$1" != "" ]; do
stripit $1
shift
done
else
while read -t 2; do
stripit $REPLY
done
fi
rm $MYTEMP
=================================================
Fix bug bug0088 reported by Mike Pont from NAG on the forum
(see http://icl.cs.utk.edu/lapack-forum/viewtopic.php?f=13&t=2893)
Actually there were a lot of problems regarding arguments checking.
I tried to correct most of them.
Apply the fix propose to all x[he/sy]rfsx.f routines
- Use IGNORE_CWISE as suggested to prevent use of unitialize variable PARAMS
I also appied the previous TESTING fix to all routines
INFO has new value in ZHERFSX (see description of INFO between ZHESVX and ZHESVXX)
This is set on line 634 (IF ( INFO .LE. N ) INFO = N + J) of zherfsx.f
And this is not handled by the testing LIN/zdrvhex.f
I just add .AND. INFO.LE.N at line 638 to avoid raising an error when INFO = N + J
At the moment, I would recommand a further look at those routines.
ZHE, ZSY led to 182 Tests failing to pass the threshold
and the same for complex
better than before, but still....
Use \par instead of \details for section.
add a Contributors Section and a Reference Section.
Remove (some) verbatim section when not needed.
Those changes have been done by hand so I am not sure I manage to catch them all.
Mark Gates