2
0
mirror of https://github.com/boostorg/ublas.git synced 2026-01-24 06:22:13 +00:00
Files
ublas/test/tensor/test_einstein_notation.cpp
Cem Bassoy 231ba5f730 refactor(core): simplify and eliminate auxiliary tensor types (#115)
Auxiliary functions for extents and strides were using different
functions. Additionally, many tags were used to distinguish between
different tensor types. This patch simplifies interfaces of different
core functions and unifies functions that can process different types of
extent and stride types.
2021-09-09 11:34:14 +02:00

147 lines
4.3 KiB
C++

//
// Copyright (c) 2018, Cem Bassoy, cem.bassoy@gmail.com
// Copyright (c) 2019, Amit Singh, amitsingh19975@gmail.com
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
// The authors gratefully acknowledge the support of
// Google and Fraunhofer IOSB, Ettlingen, Germany
//
// And we acknowledge the support from all contributors.
#include <boost/numeric/ublas/tensor.hpp>
#include <boost/test/unit_test.hpp>
#include <algorithm>
#include <iostream>
#include "utility.hpp"
BOOST_AUTO_TEST_SUITE ( test_einstein_notation/*,
*boost::unit_test::depends_on("test_multi_index") */)
using test_types = zip<int,float,std::complex<float>>::with_t<boost::numeric::ublas::layout::first_order, boost::numeric::ublas::layout::last_order>;
//using test_types = zip<int>::with_t<boost::numeric::ublas::layout::first_order>;
BOOST_AUTO_TEST_CASE_TEMPLATE( test_einstein_multiplication, value, test_types )
{
namespace ublas = boost::numeric::ublas;
using value_t = typename value::first_type;
using layout_t = typename value::second_type;
using tensor_t = ublas::tensor_dynamic<value_t,layout_t>;
// NOLINTNEXTLINE(google-build-using-namespace)
using namespace boost::numeric::ublas::index;
{
auto A = tensor_t(5,3);
auto B = tensor_t{3,4};
// auto C = tensor_t{4,5,6};
for(auto j = 0u; j < A.extents().at(1); ++j){
for(auto i = 0u; i < A.extents().at(0); ++i){
A.at( i,j ) = value_t( static_cast< inner_type_t<value_t> >(i+1) );
}
}
for(auto j = 0u; j < B.extents().at(1); ++j){
for(auto i = 0u; i < B.extents().at(0); ++i){
B.at( i,j ) = value_t( static_cast< inner_type_t<value_t> >(i+1) );
}
}
auto AB = A(_,_e) * B(_e,_);
// std::cout << "A = " << A << std::endl;
// std::cout << "B = " << B << std::endl;
// std::cout << "AB = " << AB << std::endl;
for(auto j = 0u; j < AB.extents().at(1); ++j){
for(auto i = 0u; i < AB.extents().at(0); ++i){
auto e0 = B.extents().at(0);
auto sum = std::div(e0*(e0+1),2);
auto quot = value_t(sum.quot);
BOOST_CHECK_EQUAL( AB.at(i,j) , A.at(i,0)*quot );
}
}
}
{
auto A = tensor_t{4,5,3};
auto B = tensor_t{3,4,2};
for(auto k = 0u; k < A.extents().at(2); ++k){
for(auto j = 0u; j < A.extents().at(1); ++j){
for(auto i = 0u; i < A.extents().at(0); ++i){
A.at( i,j,k ) = value_t( static_cast< inner_type_t<value_t> >(i+1) );
}
}
}
for(auto k = 0u; k < B.extents().at(2); ++k){
for(auto j = 0u; j < B.extents().at(1); ++j){
for(auto i = 0u; i < B.extents().at(0); ++i){
B.at( i,j,k ) = value_t( static_cast< inner_type_t<value_t> >(i+1) );
}
}
}
auto AB = A(_d,_,_f) * B(_f,_d,_);
// std::cout << "A = " << A << std::endl;
// std::cout << "B = " << B << std::endl;
// std::cout << "AB = " << AB << std::endl;
// n*(n+1)/2;
auto const nf = ( B.extents().at(0) * (B.extents().at(0)+1) / 2 );
auto const nd = ( A.extents().at(0) * (A.extents().at(0)+1) / 2 );
for(auto j = 0u; j < AB.extents().at(1); ++j){
for(auto i = 0u; i < AB.extents().at(0); ++i){
BOOST_CHECK_EQUAL( AB.at( i,j ) , value_t( static_cast< inner_type_t<value_t> >(nf * nd) ) );
}
}
}
{
auto A = tensor_t{{4,3}};
auto B = tensor_t{3,4,2};
for(auto j = 0u; j < A.extents().at(1); ++j){
for(auto i = 0u; i < A.extents().at(0); ++i){
A.at( i,j ) = value_t( static_cast< inner_type_t<value_t> >(i+1) );
}
}
for(auto k = 0u; k < B.extents().at(2); ++k){
for(auto j = 0u; j < B.extents().at(1); ++j){
for(auto i = 0u; i < B.extents().at(0); ++i){
B.at( i,j,k ) = value_t( static_cast< inner_type_t<value_t> >(i+1) );
}
}
}
auto AB = A(_d,_f) * B(_f,_d,_);
// n*(n+1)/2;
auto const nf = ( B.extents().at(0) * (B.extents().at(0)+1) / 2 );
auto const nd = ( A.extents().at(0) * (A.extents().at(0)+1) / 2 );
for(auto i = 0u; i < AB.extents().at(0); ++i){
BOOST_CHECK_EQUAL ( AB.at( i ) , value_t( static_cast< inner_type_t<value_t> >(nf * nd) ) );
}
}
}
BOOST_AUTO_TEST_SUITE_END()