Detailed Description

template<typename High_>
class thrill::common::UIntPair< High_ >

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 UIntPair are generally done by transforming everything to 64-bit data type, so that 64-bit register arithmetic can be used. The exception here is increment and decrement, which is done directly on the lower/higher part. Not all arithmetic operations are supported, patches welcome if you really need the operations.

Definition at line 51 of file uint_types.hpp.

#include <uint_types.hpp>

Public Types
using	High = High_
	higher part type, currently either 8-bit or 16-bit More...

using	Low = uint32_t
	lower part type, always 32-bit More...

Public Member Functions
	UIntPair ()=default
	empty constructor, does not even initialize to zero! More...

	UIntPair (const Low &l, const High &h)
	construct unit pair from lower and higher parts. More...

	UIntPair (const UIntPair &)=default
	copy constructor More...

	UIntPair (UIntPair &&)=default
	move constructor More...

	UIntPair (const uint32_t &a)
	const from a simple 32-bit unsigned integer More...

	UIntPair (const int32_t &a)
	const from a simple 32-bit signed integer More...

	UIntPair (const unsigned long long &a)
	construct from an 64-bit unsigned integer More...

	UIntPair (const unsigned long &a)
	construct from an 64-bit signed integer More...

	UIntPair (const int64_t &a)
	construct from an 64-bit signed integer More...

	operator uint64_t () const
	implicit cast to an unsigned long long More...

bool	operator!= (const UIntPair &b) const
	inequality checking operator More...

UIntPair	operator+ (const UIntPair &b) const
	addition operator (uses 64-bit arithmetic) More...

UIntPair &	operator++ ()
	prefix increment operator (directly manipulates the integer parts) More...

UIntPair &	operator+= (const UIntPair &b)
	addition operator (uses 64-bit arithmetic) More...

UIntPair	operator- (const UIntPair &b) const
	subtraction operator (uses 64-bit arithmetic) More...

UIntPair &	operator-- ()
	prefix decrement operator (directly manipulates the integer parts) More...

UIntPair &	operator-= (const UIntPair &b)
	subtraction operator (uses 64-bit arithmetic) More...

bool	operator< (const UIntPair &b) const
	less-than comparison operator More...

bool	operator< (const uint64_t &b) const
	less-than comparison operator More...

bool	operator<= (const UIntPair &b) const
	less-or-equal comparison operator More...

UIntPair &	operator= (const UIntPair &)=default
	copy assignment operator More...

UIntPair &	operator= (UIntPair &&)=default
	move assignment operator More...

bool	operator== (const UIntPair &b) const
	equality checking operator More...

bool	operator> (const UIntPair &b) const
	greater comparison operator More...

bool	operator>= (const UIntPair &b) const
	greater-or-equal comparison operator More...

uint64_t	u64 () const
	return the number as a uint64_t More...

uint64_t	ull () const
	return the number as an uint64 (unsigned long long) More...

Static Public Member Functions
static UIntPair	max ()
	return an UIntPair instance containing the largest value possible More...

static UIntPair	min ()
	return an UIntPair instance containing the smallest value possible More...

Static Public Attributes
static constexpr size_t	bytes = sizeof(Low) + sizeof(High)
	number of bytes in UIntPair More...

static constexpr size_t	digits = low_bits + high_bits
	number of binary digits (bits) in UIntPair More...

Static Private Member Functions
static unsigned	high_max ()
	return highest value storable in higher part, also used as a mask. More...

static unsigned	low_max ()
	return highest value storable in lower part, also used as a mask. More...

Private Attributes
High	high_
	member containing higher significant integer value More...

Low	low_
	member containing lower significant integer value More...

Static Private Attributes
static constexpr size_t	high_bits = 8 * sizeof(High)
	number of bits in the higher integer part, used a bit shift value. More...

static constexpr size_t	low_bits = 8 * sizeof(Low)
	number of bits in the lower integer part, used a bit shift value. More...

Friends
std::ostream &	operator<< (std::ostream &os, const UIntPair &a)
	make a UIntPair outputtable via iostreams, using unsigned long long. More...

Member Typedef Documentation

◆ High

using High = High_

higher part type, currently either 8-bit or 16-bit

Definition at line 57 of file uint_types.hpp.

◆ Low

using Low = uint32_t

lower part type, always 32-bit

Definition at line 55 of file uint_types.hpp.

Constructor & Destructor Documentation

◆ UIntPair() [1/9]

UIntPair ( )

default

empty constructor, does not even initialize to zero!

Referenced by numeric_limits< thrill::common::UIntPair< HighType > >::epsilon(), UIntPair< High_ >::max(), UIntPair< High_ >::min(), UIntPair< High_ >::operator+(), UIntPair< High_ >::operator-(), numeric_limits< thrill::common::UIntPair< HighType > >::round_error(), and UIntPair< High_ >::UIntPair().

◆ UIntPair() [2/9]

UIntPair	(	const Low &	l,
		const High &	h
	)

inline

construct unit pair from lower and higher parts.

Definition at line 96 of file uint_types.hpp.

References UIntPair< High_ >::UIntPair().

◆ UIntPair() [3/9]

UIntPair ( const UIntPair< High_ > & )

default

copy constructor

◆ UIntPair() [4/9]

UIntPair ( UIntPair< High_ > && )

default

move constructor

◆ UIntPair() [5/9]

UIntPair ( const uint32_t & a )

inline

const from a simple 32-bit unsigned integer

Definition at line 105 of file uint_types.hpp.

◆ UIntPair() [6/9]

UIntPair ( const int32_t & a )

inline

const from a simple 32-bit signed integer

Definition at line 109 of file uint_types.hpp.

References UIntPair< High_ >::high_max().

◆ UIntPair() [7/9]

UIntPair ( const unsigned long long & a )

inline

construct from an 64-bit unsigned integer

Definition at line 118 of file uint_types.hpp.

◆ UIntPair() [8/9]

UIntPair ( const unsigned long & a )

inline

construct from an 64-bit signed integer

Definition at line 126 of file uint_types.hpp.

◆ UIntPair() [9/9]

UIntPair ( const int64_t & a )

inline

construct from an 64-bit signed integer

Definition at line 130 of file uint_types.hpp.

References UIntPair< High_ >::operator=().

Member Function Documentation

◆ high_max()

static unsigned high_max ( )

inlinestaticprivate

return highest value storable in higher part, also used as a mask.

Definition at line 74 of file uint_types.hpp.

References max().

Referenced by UIntPair< High_ >::operator+(), UIntPair< High_ >::operator+=(), UIntPair< High_ >::operator-(), UIntPair< High_ >::operator-=(), and UIntPair< High_ >::UIntPair().

◆ low_max()

static unsigned low_max ( )

inlinestaticprivate

return highest value storable in lower part, also used as a mask.

Definition at line 66 of file uint_types.hpp.

References max().

Referenced by UIntPair< High_ >::operator+(), UIntPair< High_ >::operator++(), UIntPair< High_ >::operator+=(), UIntPair< High_ >::operator-(), UIntPair< High_ >::operator--(), and UIntPair< High_ >::operator-=().

◆ max()

static UIntPair max ( )

inlinestatic

return an UIntPair instance containing the largest value possible

Definition at line 250 of file uint_types.hpp.

References max(), thrill::common::TLX_ATTRIBUTE_PACKED, and UIntPair< High_ >::UIntPair().

Referenced by numeric_limits< thrill::common::UIntPair< HighType > >::max().

◆ min()

static UIntPair min ( )

inlinestatic

return an UIntPair instance containing the smallest value possible

Definition at line 244 of file uint_types.hpp.

References min(), and UIntPair< High_ >::UIntPair().

Referenced by numeric_limits< thrill::common::UIntPair< HighType > >::lowest(), and numeric_limits< thrill::common::UIntPair< HighType > >::min().

◆ operator uint64_t()

operator uint64_t ( ) const

inline

implicit cast to an unsigned long long

Definition at line 144 of file uint_types.hpp.

References UIntPair< High_ >::ull().

◆ operator!=()

bool operator!= ( const UIntPair< High_ > & b ) const

inline

inequality checking operator

Definition at line 209 of file uint_types.hpp.

References UIntPair< High_ >::high_, and UIntPair< High_ >::low_.

◆ operator+()

UIntPair operator+ ( const UIntPair< High_ > & b ) const

inline

addition operator (uses 64-bit arithmetic)

Definition at line 180 of file uint_types.hpp.

References UIntPair< High_ >::high_, UIntPair< High_ >::high_max(), UIntPair< High_ >::low_, UIntPair< High_ >::low_max(), and UIntPair< High_ >::UIntPair().

◆ operator++()

UIntPair& operator++ ( )

inline

prefix increment operator (directly manipulates the integer parts)

Definition at line 154 of file uint_types.hpp.

References UIntPair< High_ >::high_, UIntPair< High_ >::low_, UIntPair< High_ >::low_max(), and TLX_UNLIKELY.

◆ operator+=()

UIntPair& operator+= ( const UIntPair< High_ > & b )

inline

addition operator (uses 64-bit arithmetic)

Definition at line 172 of file uint_types.hpp.

References UIntPair< High_ >::high_, UIntPair< High_ >::high_max(), UIntPair< High_ >::low_, and UIntPair< High_ >::low_max().

◆ operator-()

UIntPair operator- ( const UIntPair< High_ > & b ) const

inline

subtraction operator (uses 64-bit arithmetic)

Definition at line 196 of file uint_types.hpp.

References UIntPair< High_ >::high_, UIntPair< High_ >::high_max(), UIntPair< High_ >::low_, UIntPair< High_ >::low_max(), and UIntPair< High_ >::UIntPair().

◆ operator--()

UIntPair& operator-- ( )

inline

prefix decrement operator (directly manipulates the integer parts)

Definition at line 163 of file uint_types.hpp.

References UIntPair< High_ >::high_, UIntPair< High_ >::low_, UIntPair< High_ >::low_max(), and TLX_UNLIKELY.

◆ operator-=()

UIntPair& operator-= ( const UIntPair< High_ > & b )

inline

subtraction operator (uses 64-bit arithmetic)

Definition at line 188 of file uint_types.hpp.

References UIntPair< High_ >::high_, UIntPair< High_ >::high_max(), UIntPair< High_ >::low_, and UIntPair< High_ >::low_max().

◆ operator<() [1/2]

bool operator< ( const UIntPair< High_ > & b ) const

inline

less-than comparison operator

Definition at line 214 of file uint_types.hpp.

References UIntPair< High_ >::high_, and UIntPair< High_ >::low_.

◆ operator<() [2/2]

bool operator< ( const uint64_t & b ) const

inline

less-than comparison operator

Definition at line 219 of file uint_types.hpp.

References UIntPair< High_ >::ull().

◆ operator<=()

bool operator<= ( const UIntPair< High_ > & b ) const

inline

less-or-equal comparison operator

Definition at line 224 of file uint_types.hpp.

References UIntPair< High_ >::high_, and UIntPair< High_ >::low_.

◆ operator=() [1/2]

UIntPair& operator= ( const UIntPair< High_ > & )

default

copy assignment operator

Referenced by UIntPair< High_ >::UIntPair().

◆ operator=() [2/2]

UIntPair& operator= ( UIntPair< High_ > && )

default

move assignment operator

◆ operator==()

bool operator== ( const UIntPair< High_ > & b ) const

inline

equality checking operator

Definition at line 204 of file uint_types.hpp.

References UIntPair< High_ >::high_, and UIntPair< High_ >::low_.

◆ operator>()

bool operator> ( const UIntPair< High_ > & b ) const

inline

greater comparison operator

Definition at line 229 of file uint_types.hpp.

References UIntPair< High_ >::high_, and UIntPair< High_ >::low_.

◆ operator>=()

bool operator>= ( const UIntPair< High_ > & b ) const

inline

greater-or-equal comparison operator

Definition at line 234 of file uint_types.hpp.

References UIntPair< High_ >::high_, and UIntPair< High_ >::low_.

◆ u64()

uint64_t u64 ( ) const

inline

return the number as a uint64_t

Definition at line 149 of file uint_types.hpp.

◆ ull()

uint64_t ull ( ) const

inline

return the number as an uint64 (unsigned long long)

Definition at line 139 of file uint_types.hpp.

Referenced by UIntPair< High_ >::operator uint64_t(), and UIntPair< High_ >::operator<().

Friends And Related Function Documentation

◆ operator<<

std::ostream& operator<<	(	std::ostream &	os,
		const UIntPair< High_ > &	a
	)

friend

make a UIntPair outputtable via iostreams, using unsigned long long.

Definition at line 239 of file uint_types.hpp.

Member Data Documentation

◆ bytes

constexpr size_t bytes = sizeof(Low) + sizeof(High)

static

number of bytes in UIntPair

Definition at line 86 of file uint_types.hpp.

◆ digits

constexpr size_t digits = low_bits + high_bits

static

number of binary digits (bits) in UIntPair

Definition at line 83 of file uint_types.hpp.

◆ high_

High high_

private

member containing higher significant integer value

Definition at line 63 of file uint_types.hpp.

Referenced by UIntPair< High_ >::operator!=(), UIntPair< High_ >::operator+(), UIntPair< High_ >::operator++(), UIntPair< High_ >::operator+=(), UIntPair< High_ >::operator-(), UIntPair< High_ >::operator--(), UIntPair< High_ >::operator-=(), UIntPair< High_ >::operator<(), UIntPair< High_ >::operator<=(), UIntPair< High_ >::operator==(), UIntPair< High_ >::operator>(), and UIntPair< High_ >::operator>=().

◆ high_bits

constexpr size_t high_bits = 8 * sizeof(High)

staticprivate

number of bits in the higher integer part, used a bit shift value.

Definition at line 79 of file uint_types.hpp.

◆ low_

Low low_

private

member containing lower significant integer value

Definition at line 61 of file uint_types.hpp.

Referenced by UIntPair< High_ >::operator!=(), UIntPair< High_ >::operator+(), UIntPair< High_ >::operator++(), UIntPair< High_ >::operator+=(), UIntPair< High_ >::operator-(), UIntPair< High_ >::operator--(), UIntPair< High_ >::operator-=(), UIntPair< High_ >::operator<(), UIntPair< High_ >::operator<=(), UIntPair< High_ >::operator==(), UIntPair< High_ >::operator>(), and UIntPair< High_ >::operator>=().

◆ low_bits

constexpr size_t low_bits = 8 * sizeof(Low)

staticprivate

number of bits in the lower integer part, used a bit shift value.

Definition at line 71 of file uint_types.hpp.

The documentation for this class was generated from the following file:

thrill/common/uint_types.hpp

Detailed Description

template<typename High_> class thrill::common::UIntPair< High_ >

Public Types

Public Member Functions

Static Public Member Functions

Static Public Attributes

Static Private Member Functions

Private Attributes

Static Private Attributes

Friends

Member Typedef Documentation

◆ High

◆ Low

Constructor & Destructor Documentation

◆ UIntPair() [1/9]

◆ UIntPair() [2/9]

◆ UIntPair() [3/9]

◆ UIntPair() [4/9]

◆ UIntPair() [5/9]

◆ UIntPair() [6/9]

◆ UIntPair() [7/9]

◆ UIntPair() [8/9]

◆ UIntPair() [9/9]

Member Function Documentation

◆ high_max()

◆ low_max()

◆ max()

◆ min()

◆ operator uint64_t()

◆ operator!=()

◆ operator+()

◆ operator++()

◆ operator+=()

◆ operator-()

◆ operator--()

◆ operator-=()

◆ operator<() [1/2]

◆ operator<() [2/2]

◆ operator<=()

◆ operator=() [1/2]

◆ operator=() [2/2]

◆ operator==()

◆ operator>()

◆ operator>=()

◆ u64()

◆ ull()

Friends And Related Function Documentation

◆ operator<<

Member Data Documentation

◆ bytes

◆ digits

◆ high_

◆ high_bits

◆ low_

◆ low_bits

template<typename High_>
class thrill::common::UIntPair< High_ >