TMatrix


class description - source file - inheritance tree

class TMatrix : public TObject


    protected:
void Allocate(Int_t nrows, Int_t ncols, Int_t row_lwb = 0, Int_t col_lwb = 0) void AMultB(const TMatrix& a, const TMatrix& b) void AtMultB(const TMatrix& a, const TMatrix& b) void Invalidate() void Invert(const TMatrix& m) void Transpose(const TMatrix& m) public:
TMatrix TMatrix(Int_t nrows, Int_t ncols) TMatrix TMatrix(Int_t row_lwb, Int_t row_upb, Int_t col_lwb, Int_t col_upb) TMatrix TMatrix(const TMatrix& another) TMatrix TMatrix(TMatrix::EMatrixCreatorsOp1 op, const TMatrix& prototype) TMatrix TMatrix(const TMatrix& a, TMatrix::EMatrixCreatorsOp2 op, const TMatrix& b) TMatrix TMatrix() TMatrix TMatrix(const TLazyMatrix& lazy_constructor) virtual void ~TMatrix() TMatrix& Abs() TMatrix& Apply(TElementPosAction& action) TMatrix& Apply(TElementAction& action) TClass* Class() Double_t ColNorm() const Double_t Determinant() const Double_t E2Norm() const Int_t GetColLwb() const Int_t GetColUpb() const Int_t GetNcols() const Int_t GetNoElements() const Int_t GetNrows() const Int_t GetRowLwb() const Int_t GetRowUpb() const TMatrix& HilbertMatrix() TMatrix& Invert(Double_t* determ_ptr = 0) virtual TClass* IsA() const Bool_t IsValid() const void Mult(const TMatrix& a, const TMatrix& b) Double_t Norm1() const Double_t NormInf() const Bool_t operator!=(Real_t val) const const Real_t& operator()(Int_t rown, Int_t coln) const TMatrix& operator*=(const TMatrixDiag& diag) TMatrix& operator*=(const TMatrix& source) TMatrix& operator*=(Double_t val) TMatrix& operator+=(Double_t val) TMatrix& operator-=(Double_t val) Bool_t operator<(Real_t val) const Bool_t operator<=(Real_t val) const TMatrix& operator=(const TMatrix& source) TMatrix& operator=(const TLazyMatrix& source) TMatrix& operator=(Real_t val) Bool_t operator==(Real_t val) const Bool_t operator>(Real_t val) const Bool_t operator>=(Real_t val) const virtual void Print(Option_t* option) void ResizeTo(Int_t row_lwb, Int_t row_upb, Int_t col_lwb, Int_t col_upb) void ResizeTo(const TMatrix& m) void ResizeTo(Int_t nrows, Int_t ncols) Double_t RowNorm() const virtual void ShowMembers(TMemberInspector& insp, char* parent) TMatrix& Sqr() TMatrix& Sqrt() virtual void Streamer(TBuffer& b) TMatrix& UnitMatrix() TMatrix& Zero()

Data Members

protected:
Int_t fNrows number of rows Int_t fNcols number of columns Int_t fNelems number of elements in matrix Int_t fRowLwb lower bound of the row index Int_t fColLwb lower bound of the col index Real_t* fElements elements themselves Real_t** fIndex index[i] = &matrix(0,i) (col index) public:
static const TMatrix::EMatrixCreatorsOp1 kZero static const TMatrix::EMatrixCreatorsOp1 kUnit static const TMatrix::EMatrixCreatorsOp1 kTransposed static const TMatrix::EMatrixCreatorsOp1 kInverted static const TMatrix::EMatrixCreatorsOp2 kMult static const TMatrix::EMatrixCreatorsOp2 kTransposeMult static const TMatrix::EMatrixCreatorsOp2 kInvMult static const TMatrix::EMatrixCreatorsOp2 kAtBA

Class Description

                                                                      
 Linear Algebra Package                                               
                                                                      
 The present package implements all the basic algorithms dealing      
 with vectors, matrices, matrix columns, rows, diagonals, etc.        
                                                                      
 Matrix elements are arranged in memory in a COLUMN-wise              
 fashion (in FORTRAN's spirit). In fact, it makes it very easy to     
 feed the matrices to FORTRAN procedures, which implement more        
 elaborate algorithms.                                                
                                                                      
 Unless otherwise specified, matrix and vector indices always start   
 with 0, spanning up to the specified limit-1. However, there are     
 constructors to which one can specify aribtrary lower and upper      
 bounds, e.g. TMatrix m(1,10,1,5) defines a matrix that ranges, and   
 that can be addresses, from 1..10, 1..5 (a(1,1)..a(10,5)).           
                                                                      
 The present package provides all facilities to completely AVOID      
 returning matrices. Use "TMatrix A(TMatrix::kTransposed,B);" and     
 other fancy constructors as much as possible. If one really needs    
 to return a matrix, return a TLazyMatrix object instead. The         
 conversion is completely transparent to the end user, e.g.           
 "TMatrix m = THaarMatrix(5);" and _is_ efficient.                    
                                                                      
 Since TMatrix et al. are fully integrated in ROOT they of course     
 can be stored in a ROOT database.                                    
                                                                      
                                                                      
 How to efficiently use this package                                  
 -----------------------------------                                  
                                                                      
 1. Never return complex objects (matrices or vectors)                
    Danger: For example, when the following snippet:                  
        TMatrix foo(int n)                                            
        {                                                             
           TMatrix foom(n,n); fill_in(foom); return foom;             
        }                                                             
        TMatrix m = foo(5);                                           
    runs, it constructs matrix foo:foom, copies it onto stack as a    
    return value and destroys foo:foom. Return value (a matrix)       
    from foo() is then copied over to m (via a copy constructor),     
    and the return value is destroyed. So, the matrix constructor is  
    called 3 times and the destructor 2 times. For big matrices,      
    the cost of multiple constructing/copying/destroying of objects   
    may be very large. *Some* optimized compilers can cut down on 1   
    copying/destroying, but still it leaves at least two calls to     
    the constructor. Note, TLazyMatrices (see below) can construct    
    TMatrix m "inplace", with only a _single_ call to the             
    constructor.                                                      
                                                                      
 2. Use "two-address instructions"                                    
        "void TMatrix::operator += (const TMatrix &B);"               
    as much as possible.                                              
    That is, to add two matrices, it's much more efficient to write   
        A += B;                                                       
    than                                                              
        TMatrix C = A + B;                                            
    (if both operand should be preserved,                             
        TMatrix C = A; C += B;                                        
    is still better).                                                 
                                                                      
 3. Use glorified constructors when returning of an object seems      
    inevitable:                                                       
        "TMatrix A(TMatrix::kTransposed,B);"                          
        "TMatrix C(A,TMatrix::kTransposeMult,B);"                     
                                                                      
    like in the following snippet (from $ROOTSYS/test/vmatrix.cxx)    
    that verifies that for an orthogonal matrix T, T'T = TT' = E.     
                                                                      
    TMatrix haar = THaarMatrix(5);                                    
    TMatrix unit(TMatrix::kUnit,haar);                                
    TMatrix haar_t(TMatrix::kTransposed,haar);                        
    TMatrix hth(haar,TMatrix::kTransposeMult,haar);                   
    TMatrix hht(haar,TMatrix::kMult,haar_t);                          
    TMatrix hht1 = haar; hht1 *= haar_t;                              
    VerifyMatrixIdentity(unit,hth);                                   
    VerifyMatrixIdentity(unit,hht);                                   
    VerifyMatrixIdentity(unit,hht1);                                  
                                                                      
 4. Accessing row/col/diagonal of a matrix without much fuss          
    (and without moving a lot of stuff around):                       
                                                                      
        TMatrix m(n,n); TVector v(n); TMatrixDiag(m) += 4;            
        v = TMatrixRow(m,0);                                          
        TMatrixColumn m1(m,1); m1(2) = 3; // the same as m(2,1)=3;    
    Note, constructing of, say, TMatrixDiag does *not* involve any    
    copying of any elements of the source matrix.                     
                                                                      
 5. It's possible (and encouraged) to use "nested" functions          
    For example, creating of a Hilbert matrix can be done as follows: 
                                                                      
    void foo(const TMatrix &m)                                        
    {                                                                 
       TMatrix m1(TMatrix::kZero,m);                                  
       struct MakeHilbert : public TElementPosAction {                
          void Operation(Real_t &element) { element = 1./(fI+fJ-1); } 
       };                                                             
       m1.Apply(MakeHilbert());                                       
    }                                                                 
                                                                      
    of course, using a special method TMatrix::HilbertMatrix() is     
    still more optimal, but not by a whole lot. And that's right,     
    class MakeHilbert is declared *within* a function and local to    
    that function. It means one can define another MakeHilbert class  
    (within another function or outside of any function, that is, in  
    the global scope), and it still will be OK. Note, this currently  
    is not yet supported by the interpreter CINT.                     
                                                                      
    Another example is applying of a simple function to each matrix   
    element:                                                          
                                                                      
    void foo(TMatrix &m, TMatrix &m1)                                 
    {                                                                 
       typedef  double (*dfunc_t)(double);                            
       class ApplyFunction : public TElementAction {                  
          dfunc_t *fFunc;                                             
          void Operation(Real_t &element) { element=fFunc(element); } 
       public:                                                        
          ApplyFunction(dfunc_t func):fFunc(func) {}                  
       };                                                             
       m.Apply(ApplyFunction(TMath::Sin));                            
       m1.Apply(ApplyFunction(TMath::Cos));                           
    }                                                                 
                                                                      
    Validation code $ROOTSYS/test/vmatrix.cxx and vvector.cxx contain 
    a few more examples of that kind.                                 
                                                                      
 6. Lazy matrices: instead of returning an object return a "recipe"   
    how to make it. The full matrix would be rolled out only when     
    and where it's needed:                                            
       TMatrix haar = THaarMatrix(5);                                 
    THaarMatrix() is a *class*, not a simple function. However        
    similar this looks to a returning of an object (see note #1       
    above), it's dramatically different. THaarMatrix() constructs a   
    TLazyMatrix, an object of just a few bytes long. A
    "TMatrix(const TLazyMatrix &recipe)" constructor follows the      
    recipe and makes the matrix haar() right in place. No matrix      
    element is moved whatsoever!                                      
                                                                      
 The implementation is based on original code by                      
 Oleg E. Kiselyov (oleg@pobox.com).                                   
                                                                      


void Allocate(Int_t no_rows, Int_t no_cols, Int_t row_lwb, Int_t col_lwb)
 Allocate new matrix. Arguments are number of rows, columns, row
 lowerbound (0 default) and column lowerbound (0 default).

~TMatrix()
 TMatrix destructor.

void ResizeTo(Int_t nrows, Int_t ncols)
 Erase the old matrix and create a new one according to new boundaries
 with indexation starting at 0.

void ResizeTo(Int_t row_lwb, Int_t row_upb, Int_t col_lwb, Int_t col_upb)
 Erase the old matrix and create a new one according to new boudaries.

TMatrix(EMatrixCreatorsOp1 op, const TMatrix &prototype)
 Create a matrix applying a specific operation to the prototype.
 Example: TMatrix a(10,12); ...; TMatrix b(TMatrix::kTransposed, a);
 Supported operations are: kZero, kUnit, kTransposed and kInverted.

TMatrix(const TMatrix &a, EMatrixCreatorsOp2 op, const TMatrix &b)
 Create a matrix applying a specific operation to two prototypes.
 Example: TMatrix a(10,12), b(12,5); ...; TMatrix c(a, TMatrix::kMult, b);
 Supported operations are: kMult (a*b), kTransposeMult (a'*b),
 kInvMult (a^(-1)*b) and kAtBA (a'*b*a).

Double_t RowNorm() const
 Row matrix norm, MAX{ SUM{ |M(i,j)|, over j}, over i}.
 The norm is induced by the infinity vector norm.

Double_t ColNorm() const
 Column matrix norm, MAX{ SUM{ |M(i,j)|, over i}, over j}.
 The norm is induced by the 1 vector norm.

Double_t E2Norm() const
 Square of the Euclidian norm, SUM{ m(i,j)^2 }.

void Print(Option_t *)
 Print the matrix as a table of elements (zeros are printed as dots).

void Transpose(const TMatrix &prototype)
 Transpose a matrix.

void Invert(const TMatrix &m)
 Allocate new matrix and set it to inv(m).

void AMultB(const TMatrix &a, const TMatrix &b)
 General matrix multiplication. Create a matrix C such that C = A * B.
 Note, matrix C needs to be allocated.

void Mult(const TMatrix &a, const TMatrix &b)
 Compute C = A*B. The same as AMultB(), only matrix C is already
 allocated, and it is *this.

void AtMultB(const TMatrix &a, const TMatrix &b)
 Create a matrix C such that C = A' * B. In other words,
 c[i,j] = SUM{ a[k,i] * b[k,j] }. Note, matrix C needs to be allocated.

Double_t Determinant() const
 Compute the determinant of a general square matrix.
 Example: Matrix A; Double_t A.Determinant();

 Gauss-Jordan transformations of the matrix with a slight
 modification to take advantage of the *column*-wise arrangement
 of Matrix elements. Thus we eliminate matrix's columns rather than
 rows in the Gauss-Jordan transformations. Note that determinant
 is invariant to matrix transpositions.
 The matrix is copied to a special object of type TMatrixPivoting,
 where all Gauss-Jordan eliminations with full pivoting are to
 take place.

void Streamer(TBuffer &R__b)
 Stream an object of class TMatrix.

TMatrix(Int_t no_rows, Int_t no_cols)

TMatrix(Int_t row_lwb, Int_t row_upb, Int_t col_lwb, Int_t col_upb)

TMatrix(const TLazyMatrix &lazy_constructor)

Bool_t IsValid() const

TMatrix(const TMatrix &another)

void ResizeTo(const TMatrix &m)



Inline Functions


                 void Invalidate()
              TMatrix TMatrix(const TLazyMatrix& lazy_constructor)
                Int_t GetRowLwb() const
                Int_t GetRowUpb() const
                Int_t GetNrows() const
                Int_t GetColLwb() const
                Int_t GetColUpb() const
                Int_t GetNcols() const
                Int_t GetNoElements() const
        const Real_t& operator()(Int_t rown, Int_t coln) const
             TMatrix& operator=(const TMatrix& source)
             TMatrix& operator=(const TLazyMatrix& source)
             TMatrix& operator=(Real_t val)
             TMatrix& operator-=(Double_t val)
             TMatrix& operator+=(Double_t val)
             TMatrix& operator*=(Double_t val)
               Bool_t operator==(Real_t val) const
               Bool_t operator!=(Real_t val) const
               Bool_t operator<(Real_t val) const
               Bool_t operator<=(Real_t val) const
               Bool_t operator>(Real_t val) const
               Bool_t operator>=(Real_t val) const
             TMatrix& Zero()
             TMatrix& Abs()
             TMatrix& Sqr()
             TMatrix& Sqrt()
             TMatrix& Apply(TElementAction& action)
             TMatrix& Apply(TElementPosAction& action)
             TMatrix& Invert(Double_t* determ_ptr = 0)
             TMatrix& UnitMatrix()
             TMatrix& HilbertMatrix()
             TMatrix& operator*=(const TMatrix& source)
             TMatrix& operator*=(const TMatrixDiag& diag)
             Double_t NormInf() const
             Double_t Norm1() const
              TClass* Class()
              TClass* IsA() const
                 void ShowMembers(TMemberInspector& insp, char* parent)


Author: Fons Rademakers 03/11/97
Last update: 2.22/01 20/05/99 16.31.50 by Rene Brun
Copyright (c) 1995-1999, The ROOT System, All rights reserved. *


ROOT page - Class index - Top of the page

This page has been automatically generated. If you have any comments or suggestions about the page layout send a mail to ROOT support, or contact the developers with any questions or problems regarding ROOT.