Description of scaled

scaled		Programmers Guide
		Contents \| Index \| Search

Category:containers

Component type:function

Prototype

template <class Scalable, class T> Scalable::scaled_type scaled(const Scalable& A, const T& alpha) ;

Description

This function can be used to scale arguments in MTL functions. For example, to perform the vector addition operation z <- a x + b y one would do the following:

   mtl::add(scaled(x, alpha), scaled(y, beta), z);

The actual multiplication by alpha and beta is done within the algorithm, so the performance is the same as if the add() function had parameters for alpha and beta. The scaled() function can be used with any vector or matrix argument in MTL functions. Do not confuse this function with mtl::scale() which are stand-alone functions.

Definition

scaled1D.h

Requirements on types

T must be convertible to Scalable's value_type
Scalable's value_type must be a model of Ring

Preconditions

Complexity

compile time and adds a single multiplication to each element access inside of any algorithm

Example

In y_ax_y.cc:

  typedef matrix< double, 
                  rectangle<>, 
                  dense<>, 
                  row_major>::type Matrix; 
  Matrix A(3,3);
  double SCALE = - A(2,1) / A(1,1);
  add(scaled(A[0], SCALE), A[1], A[1]);

Notes

See also