Home Browse Files

Content

Generic Implementation
CBLAS Implementation

CXXBLAS Level 1: asum

asum (defined in namespace cxxblas) computes the sum of the absolute values of elements in a vector, i.e. \(\sum\limits_{i} |x_i|\).

Generic Implementation

template <typename IndexType, typename X, typename T>
void
asum(IndexType n, const X *x, IndexType incX, T &absSum);

n	`(input)` Specifies the number of elements in vector \(x\).
x	`(input)` Pointer to the first element of vector \(x\).
incX	`(input)` Specifies the increment for indexing vector \(x\).
absSum	`(output)` Contains the sum of magnitudes of real and imaginary parts of all elements of the vector.

CBLAS Implementation

If a CBLAS implementations like ATLAS, GotoBLAS, ... is available then the following optimized variants are provided.

Single Precision

template <typename IndexType>
    typename If<IndexType>::isBlasCompatibleInteger
    asum(IndexType n, const float *x, IndexType incX,
         float &absSum);

Double Precision

template <typename IndexType>
    typename If<IndexType>::isBlasCompatibleInteger
    asum(IndexType n, const double *x, IndexType incX,
         double &absSum);

Complex Single Precision

template <typename IndexType>
    typename If<IndexType>::isBlasCompatibleInteger
    asum(IndexType n, const ComplexFloat *x, IndexType incX,
         float &absSum);

Complex Double Precision

template <typename IndexType>
    typename If<IndexType>::isBlasCompatibleInteger
    asum(IndexType n, const ComplexDouble *x, IndexType incX,
         double &absSum);

Home