/usr/include/trilinos/Teuchos_ScalarTraitsCUDA.hpp is in libtrilinos-dev 10.4.0.dfsg-1ubuntu2.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 | // @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_
/*! \file Teuchos_ScalarTraitsCUDA.hpp
\brief Defines basic traits for the scalar field type, appropriate for compilation under the NVIDIA CUDA C compiler.
*/
#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_
|