uint_types.hpp

Go to the documentation of this file.

/***************************************************************************
 *  foxxll/common/uint_types.hpp
 *
 *  Class representing a 40-bit or 48-bit unsigned integer encoded in five or
 *  six bytes.
 *
 *  Part of FOXXLL. See http://foxxll.org
 *
 *  Copyright (C) 2013 Timo Bingmann <[email protected]>
 *
 *  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 FOXXLL_COMMON_UINT_TYPES_HEADER
#define FOXXLL_COMMON_UINT_TYPES_HEADER

#include <cassert>
#include <limits>
#include <ostream>

#include <foxxll/common/types.hpp>
#include <foxxll/common/utils.hpp>
#include <foxxll/config.hpp>

#include <tlx/define/likely.hpp>

namespace foxxll {

/*!
 * Construct an 40-bit or 48-bit unsigned integer stored in five or six bytes.
 *
 * The purpose of this class is to provide integers with smaller data storage
 * footprints when more than 32-bit, but less than 64-bit indexes are
 * needed. This is commonly the case for storing file offsets and indexes. Here
 * smaller types currently suffice for files < 1 TiB or < 16 TiB.
 *
 * The class combines a 32-bit integer with a HighType (either 8-bit or 16-bit)
 * to get a larger type. Only unsigned values are supported, which fits the
 * general application of file offsets.
 *
 * Calculation in uint_pair are generally done by transforming everything to
 * 64-bit data type, so that 64-bit register arithmetic can be used. The
 * exception here is \b increment and \b decrement, which is done directly on
 * the lower/higher part. Not all arithmetic operations are supported, patches
 * welcome if you really need the operations.
 */
#if FOXXLL_MSVC
#pragma pack(push, 1)
#endif
template <typename HighType>
class uint_pair
{
public:
    //! lower part type, always 32-bit
    using low_type = uint32_t;
    //! higher part type, currently either 8-bit or 16-bit
    using high_type = HighType;

private:
    //! member containing lower significant integer value
    low_type low;
    //! member containing higher significant integer value
    high_type high;

    //! return highest value storable in lower part, also used as a mask.
    static low_type low_max()
    {
        return std::numeric_limits<low_type>::max();
    }

    //! number of bits in the lower integer part, used a bit shift value.
    static const size_t low_bits = 8 * sizeof(low_type);

    //! return highest value storable in higher part, also used as a mask.
    static high_type high_max()
    {
        return std::numeric_limits<high_type>::max();
    }

    //! number of bits in the higher integer part, used a bit shift value.
    static const size_t high_bits = 8 * sizeof(high_type);

public:
    //! number of binary digits (bits) in uint_pair
    static const size_t digits = low_bits + high_bits;

    //! number of bytes in uint_pair
    static const size_t bytes = sizeof(low_type) + sizeof(high_type);

    //! empty constructor, does not even initialize to zero!
    uint_pair()
    {
        // compile-time assertions about size of low_type
        static_assert(
            8 * sizeof(low_type) == 32,
            "8 * sizeof(low_type) == 32"
        );
        // compile-time assertions about size of our data structure, this tests
        // packing of structures by the compiler
        static_assert(
            sizeof(uint_pair) == bytes,
            "sizeof(uint_pair) == bytes"
        );
        static_assert(
            sizeof(uint_pair) == digits / 8,
            "sizeof(uint_pair) == digits / 8"
        );
        static_assert(digits / 8 == bytes, "digits / 8 == bytes");
    }

    //! construct unit pair from lower and higher parts.
    uint_pair(const low_type& l, const high_type& h)
        : low(l), high(h)
    { }

    //! implicit conversion from a simple 32-bit unsigned integer
    uint_pair(const uint32_t& a) // NOLINT
        : low(a), high(0)
    { }

    //! implicit conversion from a simple 32-bit signed integer
    uint_pair(const int32_t& a) // NOLINT
        : low(a), high(0)
    {
        if (a >= 0)
            low = a;
        else
            low = a, high = high_max();
    }

    //! implicit conversion from an uint64_t (unsigned long long)
    uint_pair(const uint64_t& a) // NOLINT
        : low(static_cast<low_type>(a & low_max())),
          high(static_cast<high_type>((a >> low_bits) & high_max()))
    {
        // check for overflow
        assert((a >> (low_bits + high_bits)) == 0);
    }

    //! return the number as an uint64_t (unsigned long long)
    uint64_t ull() const
    {
        return static_cast<uint64_t>(high) << low_bits | static_cast<uint64_t>(low);
    }

    //! implicit cast to an unsigned long long
    operator uint64_t () const
    {
        return ull();
    }

    //! return the number as a uint64_t
    uint64_t u64() const
    {
        return static_cast<uint64_t>(high) << low_bits | static_cast<uint64_t>(low);
    }

    //! prefix increment operator (directly manipulates the integer parts)
    uint_pair& operator ++ ()
    {
        if (TLX_UNLIKELY(low == low_max()))
            ++high, low = 0;
        else
            ++low;
        return *this;
    }

    //! prefix decrement operator (directly manipulates the integer parts)
    uint_pair& operator -- ()
    {
        if (TLX_UNLIKELY(low == 0))
            --high, low = low_max();
        else
            --low;
        return *this;
    }

    //! addition operator (uses 64-bit arithmetic)
    uint_pair& operator += (const uint_pair& b)
    {
        uint64_t add = static_cast<uint64_t>(low) + b.low;
        low = static_cast<low_type>(add & low_max());
        high = static_cast<high_type>(high + b.high + ((add >> low_bits) & high_max()));
        return *this;
    }

    //! equality checking operator
    bool operator == (const uint_pair& b) const
    {
        return (low == b.low) && (high == b.high);
    }

    //! inequality checking operator
    bool operator != (const uint_pair& b) const
    {
        return (low != b.low) || (high != b.high);
    }

    //! less-than comparison operator
    bool operator < (const uint_pair& b) const
    {
        return (high < b.high) || (high == b.high && low < b.low);
    }

    //! less-or-equal comparison operator
    bool operator <= (const uint_pair& b) const
    {
        return (high < b.high) || (high == b.high && low <= b.low);
    }

    //! greater comparison operator
    bool operator > (const uint_pair& b) const
    {
        return (high > b.high) || (high == b.high && low > b.low);
    }

    //! greater-or-equal comparison operator
    bool operator >= (const uint_pair& b) const
    {
        return (high > b.high) || (high == b.high && low >= b.low);
    }

    //! make a uint_pair outputtable via iostreams, using unsigned long long.
    friend std::ostream& operator << (std::ostream& os, const uint_pair& a)
    {
        return os << a.ull();
    }

    //! return an uint_pair instance containing the smallest value possible
    static uint_pair min()
    {
        return uint_pair(
            std::numeric_limits<low_type>::min(),
            std::numeric_limits<high_type>::min()
        );
    }

    //! return an uint_pair instance containing the largest value possible
    static uint_pair max()
    {
        return uint_pair(
            std::numeric_limits<low_type>::max(),
            std::numeric_limits<high_type>::max()
        );
    }
}
#if FOXXLL_MSVC
;                         // NOLINT
#pragma pack(pop)
#else
__attribute__ ((packed)); // NOLINT
#endif

//! \addtogroup foxxll_support
//! \{

//! Construct a 40-bit unsigned integer stored in five bytes.
using uint40 = uint_pair<uint8_t>;

//! Construct a 48-bit unsigned integer stored in six bytes.
using uint48 = uint_pair<uint16_t>;

//! \}

} // namespace foxxll

namespace std {

//! template class providing some numeric_limits fields for uint_pair types.
template <typename HighType>
class numeric_limits<foxxll::uint_pair<HighType> >
{
public:
    //! yes we have information about uint_pair
    static const bool is_specialized = true;

    //! return an uint_pair instance containing the smallest value possible
    static foxxll::uint_pair<HighType> min()
    { return foxxll::uint_pair<HighType>::min(); }

    //! return an uint_pair instance containing the largest value possible
    static foxxll::uint_pair<HighType> max()
    { return foxxll::uint_pair<HighType>::max(); }

    //! return an uint_pair instance containing the smallest value possible
    static foxxll::uint_pair<HighType> lowest()
    { return min(); }

    //! unit_pair types are unsigned
    static const bool is_signed = false;

    //! uint_pair types are integers
    static const bool is_integer = true;

    //! unit_pair types contain exact integers
    static const bool is_exact = true;

    //! unit_pair radix is binary
    static const int radix = 2;

    //! number of binary digits (bits) in uint_pair
    static const int digits = foxxll::uint_pair<HighType>::digits;

    //! epsilon is zero
    static const foxxll::uint_pair<HighType> epsilon()
    { return foxxll::uint_pair<HighType>(0, 0); }

    //! rounding error is zero
    static const foxxll::uint_pair<HighType> round_error()
    { return foxxll::uint_pair<HighType>(0, 0); }

    //! no exponent
    static const int min_exponent = 0;

    //! no exponent
    static const int min_exponent10 = 0;

    //! no exponent
    static const int max_exponent = 0;

    //! no exponent
    static const int max_exponent10 = 0;

    //! no infinity
    static const bool has_infinity = false;
};

} // namespace std

#endif // !FOXXLL_COMMON_UINT_TYPES_HEADER

/**************************************************************************/