Teuchos_ScalarTraitsCUDA.hpp

Go to the documentation of this file.

// @HEADER
// ***********************************************************************
//
//                    Teuchos: Common Tools Package
//                 Copyright (2004) Sandia Corporation
//
// Under terms of Contract DE-AC04-94AL85000, there is a non-exclusive
// license for use of this work by or on behalf of the U.S. Government.
//
// This library is free software; you can redistribute it and/or modify
// it under the terms of the GNU Lesser General Public License as
// published by the Free Software Foundation; either version 2.1 of the
// License, or (at your option) any later version.
//
// This library is distributed in the hope that it will be useful, but
// WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
// Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General Public
// License along with this library; if not, write to the Free Software
// Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
// USA
// Questions? Contact Michael A. Heroux (maherou@sandia.gov)
//
// ***********************************************************************
// @HEADER

#ifndef _TEUCHOS_SCALARTRAITS_CUDA_HPP_
#define _TEUCHOS_SCALARTRAITS_CUDA_HPP_

#include "Teuchos_ScalarTraitsDecl.hpp"

namespace Teuchos {

template<>
struct ScalarTraits<int>
{
  typedef int magnitudeType;
  typedef int halfPrecision;
  typedef int doublePrecision;
  static const bool isComplex = false;
  static const bool isOrdinal = true;
  static const bool isComparable = true;
  static const bool hasMachineParameters = false;
  static inline __device__ __host__ magnitudeType magnitude(int a) { return (int)fabsf((float)a); }
  static inline __device__ __host__ int zero()  { return 0; }
  static inline __device__ __host__ int one()   { return 1; }
  static inline __device__ __host__ int conjugate(int x) { return x; }
  static inline __device__ __host__ int real(int x) { return x; }
  static inline __device__ __host__ int imag(int)   { return 0; }
  static inline __device__ __host__ bool isnaninf(int) { return false; }
  static inline __device__ __host__ int squareroot(int x) { return (int)sqrtf((float)x); }          // perhaps this cast should be replaced by an explicit call like __float2int_rn
  static inline __device__ __host__ int pow(int x, int y) { return (int)powf((float)x,(float)y); }  // perhaps this cast should be replaced by an explicit call like __float2int_rn
};

#ifdef HAVE_KOKKOS_CUDA_FLOAT
template<>
struct ScalarTraits<float>
{
  typedef float magnitudeType;
  typedef float halfPrecision; // should become IEEE754-2008 binary16 or fp16 later, perhaps specified at configure according to architectural support
#ifdef HAVE_KOKKOS_CUDA_DOUBLE
  typedef double doublePrecision;
#else
  typedef float  doublePrecision;
#endif
  static const bool isComplex = false;
  static const bool isOrdinal = false;
  static const bool isComparable = true;
  static const bool hasMachineParameters = false;
  static inline __device__ __host__ magnitudeType magnitude(float a) { return fabsf(a); }
  static inline __device__ __host__ float zero()  { return(0.0f); }
  static inline __device__ __host__ float one()   { return(1.0f); }    
  static inline __device__ __host__ float conjugate(float x)   { return(x); }    
  static inline __device__ __host__ float real(float x) { return x; }
  static inline __device__ __host__ float imag(float)   { return zero(); }
  static inline __device__ __host__ bool  isnaninf(float x) { return isnan(x) || isinf(x); }
  static inline __device__ __host__ float squareroot(float x) { return sqrtf(x); }
  static inline __device__ __host__ float pow(float x, float y) { return powf(x,y); }
};
#endif // HAVE_KOKKOS_CUDA_FLOAT

#ifdef HAVE_KOKKOS_CUDA_DOUBLE
template<>
struct ScalarTraits<double>
{
  typedef double magnitudeType;
#ifdef HAVE_KOKKOS_CUDA_FLOAT
  typedef float  halfPrecision;
#else
  typedef double halfPrecision;
#endif
  typedef double doublePrecision;
  static const bool isComplex = false;
  static const bool isOrdinal = false;
  static const bool isComparable = true;
  static const bool hasMachineParameters = false;
  static inline __device__ __host__ magnitudeType magnitude(double a) { return abs(a); }
  static inline __device__ __host__ double zero()  { return(0.0); }
  static inline __device__ __host__ double one()   { return(1.0); }    
  static inline __device__ __host__ double conjugate(double x)   { return(x); }    
  static inline __device__ __host__ double real(double x) { return x; }
  static inline __device__ __host__ double imag(double)   { return zero(); }
  static inline __device__ __host__ bool  isnaninf(double x) { return isnan(x) || isinf(x); }
  static inline __device__ __host__ double squareroot(double x) { return sqrt(x); }
  static inline __device__ __host__ double pow(double x, double y) { return pow(x,y); }
};
#endif // HAVE_KOKKOS_CUDA_DOUBLE

} // Teuchos namespace

#endif // _TEUCHOS_SCALARTRAITS_CUDA_HPP_