boost/random/detail/qrng_base.hpp
/* boost random/detail/qrng_base.hpp header file
*
* Copyright Justinas Vygintas Daugmaudis 2010-2018
* 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)
*/
#ifndef BOOST_RANDOM_DETAIL_QRNG_BASE_HPP
#define BOOST_RANDOM_DETAIL_QRNG_BASE_HPP
#include <stdexcept>
#include <vector>
#include <limits>
#include <istream>
#include <ostream>
#include <sstream>
#include <boost/cstdint.hpp>
#include <boost/random/detail/operators.hpp>
#include <boost/throw_exception.hpp>
#include <boost/type_traits/integral_constant.hpp>
#include <boost/random/detail/disable_warnings.hpp>
//!\file
//!Describes the quasi-random number generator base class template.
namespace boost {
namespace random {
namespace qrng_detail {
// If the seed is a signed integer type, then we need to
// check that the value is positive:
template <typename Integer>
inline void check_seed_sign(const Integer& v, const boost::true_type)
{
if (v < 0)
{
boost::throw_exception( std::range_error("seed must be a positive integer") );
}
}
template <typename Integer>
inline void check_seed_sign(const Integer&, const boost::false_type) {}
template <typename Integer>
inline void check_seed_sign(const Integer& v)
{
check_seed_sign(v, integral_constant<bool, std::numeric_limits<Integer>::is_signed>());
}
template<typename DerivedT, typename LatticeT, typename SizeT>
class qrng_base
{
public:
typedef SizeT size_type;
typedef typename LatticeT::value_type result_type;
explicit qrng_base(std::size_t dimension)
// Guard against invalid dimensions before creating the lattice
: lattice(prevent_zero_dimension(dimension))
, quasi_state(dimension)
{
derived().seed();
}
// default copy c-tor is fine
// default assignment operator is fine
//!Returns: The dimension of of the quasi-random domain.
//!
//!Throws: nothing.
std::size_t dimension() const { return quasi_state.size(); }
//!Returns: Returns a successive element of an s-dimensional
//!(s = X::dimension()) vector at each invocation. When all elements are
//!exhausted, X::operator() begins anew with the starting element of a
//!subsequent s-dimensional vector.
//!
//!Throws: range_error.
result_type operator()()
{
return curr_elem != dimension() ? load_cached(): next_state();
}
//!Fills a range with quasi-random values.
template<typename Iter> void generate(Iter first, Iter last)
{
for (; first != last; ++first)
*first = this->operator()();
}
//!Effects: Advances *this state as if z consecutive
//!X::operator() invocations were executed.
//!
//!Throws: range_error.
void discard(boost::uintmax_t z)
{
const std::size_t dimension_value = dimension();
// Compiler knows how to optimize subsequent x / y and x % y
// statements. In fact, gcc does this even at -O1, so don't
// be tempted to "optimize" % via subtraction and multiplication.
boost::uintmax_t vec_n = z / dimension_value;
std::size_t carry = curr_elem + (z % dimension_value);
vec_n += carry / dimension_value;
carry = carry % dimension_value;
// Avoid overdiscarding by branchlessly correcting the triple
// (D, S + 1, 0) to (D, S, D) (see equality operator)
const bool corr = (!carry) & static_cast<bool>(vec_n);
// Discards vec_n (with correction) consecutive s-dimensional vectors
discard_vector(vec_n - static_cast<boost::uintmax_t>(corr));
#ifdef BOOST_MSVC
#pragma warning(push)
// disable unary minus operator applied to an unsigned type,
// result still unsigned.
#pragma warning(disable:4146)
#endif
// Sets up the proper position of the element-to-read
// curr_elem = carry + corr*dimension_value
curr_elem = carry ^ (-static_cast<std::size_t>(corr) & dimension_value);
#ifdef BOOST_MSVC
#pragma warning(pop)
#endif
}
//!Writes the textual representation of the generator to a @c std::ostream.
BOOST_RANDOM_DETAIL_OSTREAM_OPERATOR(os, qrng_base, s)
{
os << s.dimension() << " " << s.seq_count << " " << s.curr_elem;
return os;
}
//!Reads the textual representation of the generator from a @c std::istream.
BOOST_RANDOM_DETAIL_ISTREAM_OPERATOR(is, qrng_base, s)
{
std::size_t dim;
size_type seed;
boost::uintmax_t z;
if (is >> dim >> std::ws >> seed >> std::ws >> z) // initialize iff success!
{
// Check seed sign before resizing the lattice and/or recomputing state
check_seed_sign(seed);
if (s.dimension() != prevent_zero_dimension(dim))
{
s.lattice.resize(dim);
s.quasi_state.resize(dim);
}
// Fast-forward to the correct state
s.derived().seed(seed);
if (z != 0) s.discard(z);
}
return is;
}
//!Returns true if the two generators will produce identical sequences of outputs.
BOOST_RANDOM_DETAIL_EQUALITY_OPERATOR(qrng_base, x, y)
{
const std::size_t dimension_value = x.dimension();
// Note that two generators with different seq_counts and curr_elems can
// produce the same sequence because the generator triple
// (D, S, D) is equivalent to (D, S + 1, 0), where D is dimension, S -- seq_count,
// and the last one is curr_elem.
return (dimension_value == y.dimension()) &&
// |x.seq_count - y.seq_count| <= 1
!((x.seq_count < y.seq_count ? y.seq_count - x.seq_count : x.seq_count - y.seq_count)
> static_cast<size_type>(1)) &&
// Potential overflows don't matter here, since we've already ascertained
// that sequence counts differ by no more than 1, so if they overflow, they
// can overflow together.
(x.seq_count + (x.curr_elem / dimension_value) == y.seq_count + (y.curr_elem / dimension_value)) &&
(x.curr_elem % dimension_value == y.curr_elem % dimension_value);
}
//!Returns true if the two generators will produce different sequences of outputs.
BOOST_RANDOM_DETAIL_INEQUALITY_OPERATOR(qrng_base)
protected:
typedef std::vector<result_type> state_type;
typedef typename state_type::iterator state_iterator;
// Getters
size_type curr_seq() const { return seq_count; }
state_iterator state_begin() { return quasi_state.begin(); }
state_iterator state_end() { return quasi_state.end(); }
// Setters
void reset_seq(size_type seq)
{
seq_count = seq;
curr_elem = 0u;
}
private:
DerivedT& derived() throw()
{
return *static_cast<DerivedT * const>(this);
}
// Load the result from the saved state.
result_type load_cached()
{
return quasi_state[curr_elem++];
}
result_type next_state()
{
size_type new_seq = seq_count;
if (BOOST_LIKELY(++new_seq > seq_count))
{
derived().compute_seq(new_seq);
reset_seq(new_seq);
return load_cached();
}
boost::throw_exception( std::range_error("qrng_base: next_state") );
}
// Discards z consecutive s-dimensional vectors,
// and preserves the position of the element-to-read
void discard_vector(boost::uintmax_t z)
{
const boost::uintmax_t max_z = (std::numeric_limits<size_type>::max)() - seq_count;
// Don't allow seq_count + z overflows here
if (max_z < z)
boost::throw_exception( std::range_error("qrng_base: discard_vector") );
std::size_t tmp = curr_elem;
derived().seed(static_cast<size_type>(seq_count + z));
curr_elem = tmp;
}
static std::size_t prevent_zero_dimension(std::size_t dimension)
{
if (dimension == 0)
boost::throw_exception( std::invalid_argument("qrng_base: zero dimension") );
return dimension;
}
// Member variables are so ordered with the intention
// that the typical memory access pattern would be
// incremental. Moreover, lattice is put before quasi_state
// because we want to construct lattice first. Lattices
// can do some kind of dimension sanity check (as in
// dimension_assert below), and if that fails then we don't
// need to do any more work.
private:
std::size_t curr_elem;
size_type seq_count;
protected:
LatticeT lattice;
private:
state_type quasi_state;
};
inline void dimension_assert(const char* generator, std::size_t dim, std::size_t maxdim)
{
if (!dim || dim > maxdim)
{
std::ostringstream os;
os << "The " << generator << " quasi-random number generator only supports dimensions in range [1; "
<< maxdim << "], but dimension " << dim << " was supplied.";
boost::throw_exception( std::invalid_argument(os.str()) );
}
}
} // namespace qrng_detail
} // namespace random
} // namespace boost
#include <boost/random/detail/enable_warnings.hpp>
#endif // BOOST_RANDOM_DETAIL_QRNG_BASE_HPP