Log of /trunk/math/matrix/inc/TDecompBK.h
Parent Directory
Revision
22885 -
(
view)
(
download)
(
as text)
(
annotate)
-
[select for diffs]
Modified
Fri Mar 28 13:57:25 2008 UTC (6 years, 9 months ago) by
rdm
File length: 3138 byte(s)
Diff to
previous 20882
move the following directories under the new "math" meta directory:
mathcore
mathmore
fftw
foam
fumili
genvector
matrix
minuit
minuit2
mlp
physics
smatrix
splot
unuran
quadp
Revision
16458 -
(
view)
(
download)
(
as text)
(
annotate)
-
[select for diffs]
Modified
Fri Oct 6 06:52:34 2006 UTC (8 years, 3 months ago) by
brun
Original Path:
trunk/matrix/inc/TDecompBK.h
File length: 3197 byte(s)
Diff to
previous 15068
From Eddy Offermann:
Previously, the matrix package contained a large amount of ASSERT
statements which are a nuisance for programs analyzing a series of
independent events like in high-energy physics .
Assert's were issued when for instance a division by zero was requested
or a matrix was invalid . Most algorithms made matrices/vectors invalid
after an error occured in an operation , like inversion of a singular matrix .
Unfortunately, not all assert's were accompanied by error messages .
This situation has been completely overhauled :
- All error conditions in the algorithms are now accompanied by error
messages .
- In all algorithms it is still asserted that vectors/matrices are
valid BUT only in very few cases is a matrix/vector made invalid :
for instance if memory is allocated with incorrect parameters .
- In case of division by zero, the division is skipped . In case of a
singular matrix, the inversion routine returns the original matrix .
In the past the result of an inversion could be checked through the
value of the returned determinant or checking whether the inverted matrix
was valid .
Since from now on, we never make the matrix invalid in this operation, the
latter check will not indicate a singular matrix anymore .
The decompostion classes TDecomp... have a backward-compatible
change in the interface which makes detection of singularity easier :
old interface :
void Invert (TMatrixD &inv);
TMatrixD Invert ();
new interface :
Bool_t Invert (TMatrixD &inv);
TMatrixD Invert (Bool_t &status);
TMatrixD Invert () { Bool_t status; return Invert(status);
}
The returned status is kFALSE in case of singularity .
The old situation is easily reproduced by setting the ROOT variable
gErrorAbortLevel to kError . This cause an exception when there is an error
message (Error...) ., In the past the matrix would be made invalid which
would cause the next operation to throw an exception .
Revision
13751 -
(
view)
(
download)
(
as text)
(
annotate)
-
[select for diffs]
Modified
Thu Dec 22 08:41:32 2005 UTC (9 years, 1 month ago) by
brun
Original Path:
trunk/matrix/inc/TDecompBK.h
File length: 3104 byte(s)
Diff to
previous 10329
From Eddy Offermann;
Introduce a templated version of the matrix package.
WARNING: Read carefully the Release Notes of version 5.08
concerning the forward declarations of classes TMatriXF,TMatrixD,TVectorF,TVectorD. Use #include "TMatrixDfwh.h, etc" instead.
If uou update from CVS, do
cd $ROOTSYS/matrix/src
rm *.d *.o
Revision
10329 -
(
view)
(
download)
(
as text)
(
annotate)
-
[select for diffs]
Added
Sat Oct 16 18:09:17 2004 UTC (10 years, 3 months ago) by
brun
Original Path:
trunk/matrix/inc/TDecompBK.h
File length: 2997 byte(s)
From Eddy Offermann:
- added new decomposition class TDecompBK: Bunch-Kaufman algorithm
this class is designed to decompose real symmetric matrices . Now
we can guarantee that inverse of a symmetric matrix is again
symmetric .
- As a consequence of the introduction of TDecompBK, we added
Invert(..) functionality to TMatrixD/FSym
Also added InvertFast to TMatrixD/FSym which uses Cramer
inversion for matrices <= (6 x 6) , resulting in large speedup
but less accuracy . The necessary routines are encapsulated
in the new TMatrixDSymCramerInv class .
Note of caution :
Unlike the other decompositions, Bunch-Kaufmann does not result
in a triangular matrix of which the determinant is trivially calculated
by multiplying the diagonal elements . As a consequence,
we have to invoke yet another decomposition when the user
requests the determinant .
So Invert() will be roughly twice faster than Double_t *det; Invert(det)
- Changed the algoritm for the Cholesky decomposition so that we
do not need to have a copy of the original matrix around !
After removal of class member fA, increased the version to 2
- Added sorting of eigen -values/-vectors to TMatrixDEigen .
They are now sorted according to Re^2+Im^2 of the (possible)
complex eigenvalues in descending order . TMatrixDSymEigen
already sorted . Since most users use TMatrixD instead of
TMatrixDSym, this will ensure backward compatiblity since
old root contained the symmetric algorithms in TMatrixD .
- Added rank 1 update algorihms : A += alpha * x y^T
(x and y are vectors)
to the non-sparse matrix classes (adding this to sparse
would make this matrix immediately dense :)
- Moved Invert fom TDecompBase to the different decomposition
classes . Reason is that some classes can decompose A(m,n)
where m > n . The result of Invert() is then a "pseudo-invert"
: A_pseudo = (A^T A)^-1 A^T .
- Removed an unfortunate bug in the TMatrixD/FSub class :
both operators
void operator+=(Double_t val);
void operator*=(Double_t val);
performed actually
void operator=(Double_t val);
~
This form allows you to request diffs between any two revisions of this file.
For each of the two "sides" of the diff,
enter a numeric revision.
