42 #ifndef _TEUCHOS_SERIALBANDDENSEMATRIX_HPP_
43 #define _TEUCHOS_SERIALBANDDENSEMATRIX_HPP_
54 #include "Teuchos_Assert.hpp"
132 template<
typename OrdinalType,
typename ScalarType>
293 ScalarType&
operator () (OrdinalType rowIndex, OrdinalType colIndex);
303 const ScalarType&
operator () (OrdinalType rowIndex, OrdinalType colIndex)
const;
323 const ScalarType*
operator [] (OrdinalType colIndex)
const;
327 ScalarType*
values()
const {
return(values_); }
357 int scale (
const ScalarType alpha );
391 OrdinalType
numRows()
const {
return(numRows_); }
394 OrdinalType
numCols()
const {
return(numCols_); }
403 OrdinalType
stride()
const {
return(stride_); }
406 bool empty()
const {
return(numRows_ == 0 || numCols_ == 0); }
426 virtual std::ostream&
print(std::ostream& os)
const;
431 void copyMat(ScalarType* inputMatrix, OrdinalType strideInput,
432 OrdinalType
numRows, OrdinalType
numCols, ScalarType* outputMatrix,
435 void checkIndex( OrdinalType rowIndex, OrdinalType colIndex = 0 )
const;
436 OrdinalType numRows_;
437 OrdinalType numCols_;
450 template<
typename OrdinalType,
typename ScalarType>
453 BLAS<OrdinalType,ScalarType>(),
459 valuesCopied_ (false),
463 template<
typename OrdinalType,
typename ScalarType>
466 OrdinalType numCols_in,
471 BLAS<OrdinalType,ScalarType>(),
472 numRows_ (numRows_in),
473 numCols_ (numCols_in),
474 stride_ (kl_in+ku_in+1),
477 valuesCopied_ (true),
480 values_ =
new ScalarType[stride_ * numCols_];
486 template<
typename OrdinalType,
typename ScalarType>
489 ScalarType* values_in,
490 OrdinalType stride_in,
491 OrdinalType numRows_in,
492 OrdinalType numCols_in,
496 BLAS<OrdinalType,ScalarType>(),
497 numRows_ (numRows_in),
498 numCols_ (numCols_in),
502 valuesCopied_ (false),
507 values_ =
new ScalarType[stride_*numCols_];
508 copyMat (values_in, stride_in, numRows_, numCols_, values_, stride_, 0);
509 valuesCopied_ =
true;
513 template<
typename OrdinalType,
typename ScalarType>
517 BLAS<OrdinalType,ScalarType>(),
523 valuesCopied_ (true),
527 numRows_ = Source.numRows_;
528 numCols_ = Source.numCols_;
532 values_ =
new ScalarType[stride_*numCols_];
533 copyMat (Source.values_, Source.stride_, numRows_, numCols_,
534 values_, stride_, 0);
537 numRows_ = Source.numCols_;
538 numCols_ = Source.numRows_;
542 values_ =
new ScalarType[stride_*numCols_];
543 for (OrdinalType j = 0; j < numCols_; ++j) {
544 for (OrdinalType i = TEUCHOS_MAX(0,j-ku_);
545 i <= TEUCHOS_MIN(numRows_-1,j+kl_); ++i) {
546 values_[j*stride_ + (ku_+i-j)] =
552 numRows_ = Source.numCols_;
553 numCols_ = Source.numRows_;
557 values_ =
new ScalarType[stride_*numCols_];
558 for (OrdinalType j=0; j<numCols_; j++) {
559 for (OrdinalType i = TEUCHOS_MAX(0,j-ku_);
560 i <= TEUCHOS_MIN(numRows_-1,j+kl_); ++i) {
561 values_[j*stride_ + (ku_+i-j)] = Source.values_[i*Source.stride_ + (Source.ku_+j-i)];
567 template<
typename OrdinalType,
typename ScalarType>
571 OrdinalType numRows_in,
572 OrdinalType numCols_in,
573 OrdinalType startCol)
575 BLAS<OrdinalType,ScalarType>(),
576 numRows_ (numRows_in),
577 numCols_ (numCols_in),
578 stride_ (Source.stride_),
581 valuesCopied_ (false),
582 values_ (Source.values_)
585 values_ =
new ScalarType[stride_ * numCols_in];
586 copyMat (Source.values_, Source.stride_, numRows_in, numCols_in,
587 values_, stride_, startCol);
588 valuesCopied_ =
true;
590 values_ = values_ + (stride_ * startCol);
594 template<
typename OrdinalType,
typename ScalarType>
604 template<
typename OrdinalType,
typename ScalarType>
606 OrdinalType numRows_in, OrdinalType numCols_in, OrdinalType kl_in, OrdinalType ku_in
610 numRows_ = numRows_in;
611 numCols_ = numCols_in;
615 values_ =
new ScalarType[stride_*numCols_];
617 valuesCopied_ =
true;
622 template<
typename OrdinalType,
typename ScalarType>
624 OrdinalType numRows_in, OrdinalType numCols_in, OrdinalType kl_in, OrdinalType ku_in
628 numRows_ = numRows_in;
629 numCols_ = numCols_in;
633 values_ =
new ScalarType[stride_*numCols_];
634 valuesCopied_ =
true;
639 template<
typename OrdinalType,
typename ScalarType>
641 OrdinalType numRows_in, OrdinalType numCols_in, OrdinalType kl_in, OrdinalType ku_in
646 ScalarType* values_tmp =
new ScalarType[(kl_in+ku_in+1) * numCols_in];
648 for(OrdinalType k = 0; k < (kl_in+ku_in+1) * numCols_in; k++) {
649 values_tmp[k] = zero;
651 OrdinalType numRows_tmp = TEUCHOS_MIN(numRows_, numRows_in);
652 OrdinalType numCols_tmp = TEUCHOS_MIN(numCols_, numCols_in);
654 copyMat(values_, stride_, numRows_tmp, numCols_tmp, values_tmp,
658 numRows_ = numRows_in;
659 numCols_ = numCols_in;
663 values_ = values_tmp;
664 valuesCopied_ =
true;
673 template<
typename OrdinalType,
typename ScalarType>
678 for(OrdinalType j = 0; j < numCols_; j++) {
679 for (OrdinalType i=TEUCHOS_MAX(0,j-ku_); i<=TEUCHOS_MIN(numRows_-1,j+kl_); i++) {
680 values_[(ku_+i-j) + j*stride_] = value_in;
687 template<
typename OrdinalType,
typename ScalarType>
692 for(OrdinalType j = 0; j < numCols_; j++) {
693 for (OrdinalType i=TEUCHOS_MAX(0,j-ku_); i<=TEUCHOS_MIN(numRows_-1,j+kl_); i++) {
701 template<
typename OrdinalType,
typename ScalarType>
710 if((!valuesCopied_) && (!Source.valuesCopied_) && (values_ == Source.values_))
714 if (!Source.valuesCopied_) {
719 numRows_ = Source.numRows_;
720 numCols_ = Source.numCols_;
723 stride_ = Source.stride_;
724 values_ = Source.values_;
728 numRows_ = Source.numRows_;
729 numCols_ = Source.numCols_;
733 const OrdinalType newsize = stride_ * numCols_;
735 values_ =
new ScalarType[newsize];
736 valuesCopied_ =
true;
742 if((Source.numRows_ <= stride_) && (Source.numCols_ == numCols_)) {
743 numRows_ = Source.numRows_;
744 numCols_ = Source.numCols_;
750 numRows_ = Source.numRows_;
751 numCols_ = Source.numCols_;
755 const OrdinalType newsize = stride_ * numCols_;
757 values_ =
new ScalarType[newsize];
758 valuesCopied_ =
true;
762 copyMat(Source.values_, Source.stride_, numRows_, numCols_, values_, stride_, 0);
768 template<
typename OrdinalType,
typename ScalarType>
773 if ((numRows_ != Source.numRows_) || (numCols_ != Source.numCols_) || (kl_ != Source.kl_) || (ku_ != Source.ku_)) {
774 TEUCHOS_CHK_REF(*
this);
781 template<
typename OrdinalType,
typename ScalarType>
786 if ((numRows_ != Source.numRows_) || (numCols_ != Source.numCols_) || (kl_ != Source.kl_) || (ku_ != Source.ku_)) {
787 TEUCHOS_CHK_REF(*
this);
794 template<
typename OrdinalType,
typename ScalarType>
799 if((!valuesCopied_) && (!Source.valuesCopied_) && (values_ == Source.values_))
803 if ((numRows_ != Source.numRows_) || (numCols_ != Source.numCols_) || (kl_ != Source.kl_) || (ku_ != Source.ku_)) {
804 TEUCHOS_CHK_REF(*
this);
806 copyMat(Source.values_, Source.stride_, numRows_, numCols_, values_, stride_, 0);
815 template<
typename OrdinalType,
typename ScalarType>
818 #ifdef HAVE_TEUCHOS_ARRAY_BOUNDSCHECK
819 checkIndex( rowIndex, colIndex );
821 return(values_[colIndex * stride_ + ku_+rowIndex-colIndex]);
824 template<
typename OrdinalType,
typename ScalarType>
827 #ifdef HAVE_TEUCHOS_ARRAY_BOUNDSCHECK
828 checkIndex( rowIndex, colIndex );
830 return(values_[colIndex * stride_ + ku_+rowIndex-colIndex]);
833 template<
typename OrdinalType,
typename ScalarType>
836 #ifdef HAVE_TEUCHOS_ARRAY_BOUNDSCHECK
837 checkIndex( 0, colIndex );
839 return(values_ + colIndex * stride_);
842 template<
typename OrdinalType,
typename ScalarType>
845 #ifdef HAVE_TEUCHOS_ARRAY_BOUNDSCHECK
846 checkIndex( 0, colIndex );
848 return(values_ + colIndex * stride_);
855 template<
typename OrdinalType,
typename ScalarType>
863 for(j = 0; j < numCols_; j++) {
865 ptr = values_ + j * stride_ + TEUCHOS_MAX(0, ku_-j);
866 for (i=TEUCHOS_MAX(0,j-ku_); i<=TEUCHOS_MIN(numRows_-1,j+kl_); i++) {
874 updateFlops((kl_+ku_+1) * numCols_);
879 template<
typename OrdinalType,
typename ScalarType>
885 for (i = 0; i < numRows_; i++) {
887 for (j=TEUCHOS_MAX(0,i-kl_); j<=TEUCHOS_MIN(numCols_-1,i+ku_); j++) {
890 anorm = TEUCHOS_MAX( anorm, sum );
892 updateFlops((kl_+ku_+1) * numCols_);
897 template<
typename OrdinalType,
typename ScalarType>
903 for (j = 0; j < numCols_; j++) {
904 for (i=TEUCHOS_MAX(0,j-ku_); i<=TEUCHOS_MIN(numRows_-1,j+kl_); i++) {
909 updateFlops((kl_+ku_+1) * numCols_);
918 template<
typename OrdinalType,
typename ScalarType>
923 if((numRows_ != Operand.numRows_) || (numCols_ != Operand.numCols_) || (kl_ != Operand.kl_) || (ku_ != Operand.ku_)) {
927 for(j = 0; j < numCols_; j++) {
928 for (i=TEUCHOS_MAX(0,j-ku_); i<=TEUCHOS_MIN(numRows_-1,j+kl_); i++) {
929 if((*
this)(i, j) != Operand(i, j)) {
939 template<
typename OrdinalType,
typename ScalarType>
942 return !((*this) == Operand);
949 template<
typename OrdinalType,
typename ScalarType>
952 this->scale( alpha );
956 template<
typename OrdinalType,
typename ScalarType>
963 for (j=0; j<numCols_; j++) {
964 ptr = values_ + j*stride_ + TEUCHOS_MAX(0, ku_-j);
965 for (i=TEUCHOS_MAX(0,j-ku_); i<=TEUCHOS_MIN(numRows_-1,j+kl_); i++) {
966 *ptr = alpha * (*ptr); ptr++;
969 updateFlops( (kl_+ku_+1)*numCols_ );
974 template<
typename OrdinalType,
typename ScalarType>
982 if ((numRows_ != A.numRows_) || (numCols_ != A.numCols_) || (kl_ != A.kl_) || (ku_ != A.ku_)) {
985 for (j=0; j<numCols_; j++) {
986 ptr = values_ + j*stride_ + TEUCHOS_MAX(0, ku_-j);
987 for (i=TEUCHOS_MAX(0,j-ku_); i<=TEUCHOS_MIN(numRows_-1,j+kl_); i++) {
988 *ptr = A(i,j) * (*ptr); ptr++;
991 updateFlops( (kl_+ku_+1)*numCols_ );
996 template<
typename OrdinalType,
typename ScalarType>
1001 os <<
"Values_copied : yes" << std::endl;
1003 os <<
"Values_copied : no" << std::endl;
1004 os <<
"Rows : " << numRows_ << std::endl;
1005 os <<
"Columns : " << numCols_ << std::endl;
1006 os <<
"Lower Bandwidth : " << kl_ << std::endl;
1007 os <<
"Upper Bandwidth : " << ku_ << std::endl;
1008 os <<
"LDA : " << stride_ << std::endl;
1009 if(numRows_ == 0 || numCols_ == 0) {
1010 os <<
"(matrix is empty, no values to display)" << std::endl;
1013 for(OrdinalType i = 0; i < numRows_; i++) {
1014 for (OrdinalType j=TEUCHOS_MAX(0,i-kl_); j<=TEUCHOS_MIN(numCols_-1,i+ku_); j++) {
1015 os << (*this)(i,j) <<
" ";
1027 template<
typename OrdinalType,
typename ScalarType>
1031 rowIndex < TEUCHOS_MAX(0,colIndex-ku_) || rowIndex > TEUCHOS_MIN(numRows_-1,colIndex+kl_),
1033 "SerialBandDenseMatrix<T>::checkIndex: "
1034 "Row index " << rowIndex <<
" out of range [0, "<< numRows_ <<
")");
1036 "SerialBandDenseMatrix<T>::checkIndex: "
1037 "Col index " << colIndex <<
" out of range [0, "<< numCols_ <<
")");
1041 template<
typename OrdinalType,
typename ScalarType>
1042 void SerialBandDenseMatrix<OrdinalType, ScalarType>::deleteArrays(
void)
1044 if (valuesCopied_) {
1047 valuesCopied_ =
false;
1051 template<
typename OrdinalType,
typename ScalarType>
1052 void SerialBandDenseMatrix<OrdinalType, ScalarType>::copyMat(
1053 ScalarType* inputMatrix, OrdinalType strideInput, OrdinalType numRows_in,
1054 OrdinalType numCols_in, ScalarType* outputMatrix, OrdinalType strideOutput, OrdinalType startCol, ScalarType alpha
1058 ScalarType* ptr1 = 0;
1059 ScalarType* ptr2 = 0;
1061 for(j = 0; j < numCols_in; j++) {
1062 ptr1 = outputMatrix + (j * strideOutput) + TEUCHOS_MAX(0, ku_-j);
1063 ptr2 = inputMatrix + (j + startCol) * strideInput + TEUCHOS_MAX(0, ku_-j);
1065 for (i=TEUCHOS_MAX(0,j-ku_); i<=TEUCHOS_MIN(numRows_in-1,j+kl_); i++) {
1066 *ptr1++ += alpha*(*ptr2++);
1069 for (i=TEUCHOS_MAX(0,j-ku_); i<=TEUCHOS_MIN(numRows_in-1,j+kl_); i++) {
1077 template<
typename OrdinalType,
typename ScalarType>
1087 template<
typename OrdinalType,
typename ScalarType>
1089 operator<<(std::ostream &out,
1092 printer.obj.print(out);
1097 template<
typename OrdinalType,
typename ScalarType>
1098 SerialBandDenseMatrixPrinter<OrdinalType,ScalarType>
bool empty() const
Returns whether this matrix is empty.
OrdinalType numRows() const
Returns the row dimension of this matrix.
SerialBandDenseMatrix< OrdinalType, ScalarType > & operator+=(const SerialBandDenseMatrix< OrdinalType, ScalarType > &Source)
Add another matrix to this matrix.
ScalarType & operator()(OrdinalType rowIndex, OrdinalType colIndex)
Element access method (non-const).
int scale(const ScalarType alpha)
Scale this matrix by alpha; *this = alpha**this.
int shape(OrdinalType numRows, OrdinalType numCols, OrdinalType kl, OrdinalType ku)
Shape method for changing the size of a SerialBandDenseMatrix, initializing entries to zero...
ScalarTraits< ScalarType >::magnitudeType normInf() const
Returns the Infinity-norm of the matrix.
Ostream manipulator for SerialBandDenseMatrix.
Templated interface class to BLAS routines.
SerialBandDenseMatrixPrinter< OrdinalType, ScalarType > printMat(const SerialBandDenseMatrix< OrdinalType, ScalarType > &obj)
Return SerialBandDenseMatrix ostream manipulator Use as:
virtual ~SerialBandDenseMatrix()
Destructor.
#define TEUCHOS_TEST_FOR_EXCEPTION(throw_exception_test, Exception, msg)
Macro for throwing an exception with breakpointing to ease debugging.
int reshape(OrdinalType numRows, OrdinalType numCols, OrdinalType kl, OrdinalType ku)
Reshaping method for changing the size of a SerialBandDenseMatrix, keeping the entries.
Teuchos header file which uses auto-configuration information to include necessary C++ headers...
ScalarTraits< ScalarType >::magnitudeType normOne() const
Returns the 1-norm of the matrix.
SerialBandDenseMatrix< OrdinalType, ScalarType > & assign(const SerialBandDenseMatrix< OrdinalType, ScalarType > &Source)
Copies values from one matrix to another.
Object for storing data and providing functionality that is common to all computational classes...
virtual std::ostream & print(std::ostream &os) const
Print method. Defines the behavior of the std::ostream << operator.
This structure defines some basic traits for a scalar field type.
OrdinalType stride() const
Returns the stride between the columns of this matrix in memory.
ScalarType * values() const
Data array access method.
Functionality and data that is common to all computational classes.
bool operator!=(const SerialBandDenseMatrix< OrdinalType, ScalarType > &Operand) const
Inequality of two matrices.
ScalarTraits< ScalarType >::magnitudeType normFrobenius() const
Returns the Frobenius-norm of the matrix.
This class creates and provides basic support for banded dense matrices of templated type...
SerialBandDenseMatrix< OrdinalType, ScalarType > & operator*=(const ScalarType alpha)
Scale this matrix by alpha; *this = alpha**this.
SerialBandDenseMatrix< OrdinalType, ScalarType > & operator-=(const SerialBandDenseMatrix< OrdinalType, ScalarType > &Source)
Subtract another matrix from this matrix.
int putScalar(const ScalarType value=Teuchos::ScalarTraits< ScalarType >::zero())
Set all values in the matrix to a constant value.
int random()
Set all values in the matrix to be random numbers.
SerialBandDenseMatrix()
Default Constructor.
bool operator==(const SerialBandDenseMatrix< OrdinalType, ScalarType > &Operand) const
Equality of two matrices.
static magnitudeType magnitude(T a)
Returns the magnitudeType of the scalar type a.
OrdinalType ordinalType
Typedef for ordinal type.
SerialBandDenseMatrix< OrdinalType, ScalarType > & operator=(const SerialBandDenseMatrix< OrdinalType, ScalarType > &Source)
Copies values from one matrix to another.
OrdinalType upperBandwidth() const
Returns the upper bandwidth of this matrix.
Defines basic traits for the scalar field type.
static T zero()
Returns representation of zero for this scalar type.
OrdinalType lowerBandwidth() const
Returns the lower bandwidth of this matrix.
ScalarType * operator[](OrdinalType colIndex)
Column access method (non-const).
ScalarType scalarType
Typedef for scalar type.
OrdinalType numCols() const
Returns the column dimension of this matrix.
Reference-counted pointer class and non-member templated function implementations.
int shapeUninitialized(OrdinalType numRows, OrdinalType numCols, OrdinalType kl, OrdinalType ku)
Same as shape() except leaves uninitialized.
Teuchos::DataAccess Mode enumerable type.