Burger::FPLargeInt Class Reference

Class for managing a very large integer for double precision. More...

Public Types
typedef uint32_t	ChunkSize_t
	Size of each data chunk.
typedef uint64_t	MathSize_t
	Size of data chunk for math (Double size of ChunkSize_t)

Public Member Functions
	FPLargeInt (void) noexcept
	Default constructor.
	FPLargeInt (uint32_t uBitsNeeded) noexcept
	Constructor that sets to a default.
uint_t	is_zero (void) const noexcept
	Returns TRUE if the value is zero.
uint_t	is_not_zero (void) const noexcept
	Returns TRUE if the value is not zero.
void	init (uint32_t uBitsNeeded) noexcept
	Initialize the giant integer.
void	insert_bits_at_end (uint32_t uBits, uint32_t uShiftAmount) noexcept
	Insert bits into the buffer.
void	insert_bits_at_start (uint32_t uBits, uint32_t uShiftAmount) noexcept
	Insert bits into the buffer from the top.
uint32_t	divide_return_remainder (uint32_t uDivisor) noexcept
	Divide the long number and return the remainder.
uint32_t	multiply_return_overflow (uint32_t uMultiplier) noexcept
	Multiply by multiplier and return the overflow, from 0 to div-1.

Static Public Attributes
static constexpr const uint32_t	kBitsPerElement
	Number of bits per uint32_t.
static constexpr const uint32_t	kTotalBitsInTable = 1024 + 64
	Number of bits in the table to handle double precision.
static constexpr const uint32_t	kMaxElements
	Number of elements in the data table.

Protected Member Functions
void	insert_chunk_bits (ChunkSize_t uBits, uint32_t uEntryIndex) noexcept
	Do the actual work of inserting bits, and updating zero flag, and extents.

Protected Attributes
uint32_t	m_uEntryCount
	Number of entries in the data table.
int32_t	m_iHighestNonZeroElement
	Highest index with a non-zero entry in the data table.
int32_t	m_iLowestNonZeroElement
	Lowest index with a non-zero entry in the data table.
uint32_t	m_bIsZero
	TRUE if a zero length buffer
ChunkSize_t	m_uDataTable [kMaxElements]
	Data table of bits for locating digits.

Detailed Description

Class for managing a very large integer for double precision.

---

This class is useful in parsing high precision numbers and can help in decoding floating point numbers for eventual conversion into a string. A double has up to 2048+53 bits of precision with half in the integer range and the other half in the fractional range. This class handles 1024+64 bits to be able to handle the entire fractional or integer range for ultra high precision double to ASCII conversion.

It can perform modulo and multiplication across all bits of precision. Since this class was designed for high precision integer to ASCII conversion, the two main functions are multiply_return_overflow() to parse out digits from the fractional side of a fixed point number and divide_return_remainder() to extract digits from the integer side of the fixed point number.

Member Typedef Documentation

ChunkSize_t

uint32_t Burger::FPLargeInt::ChunkSize_t

Size of each data chunk.

MathSize_t

uint64_t Burger::FPLargeInt::MathSize_t

Size of data chunk for math (Double size of ChunkSize_t)

Constructor & Destructor Documentation

FPLargeInt() [1/2]

Burger::FPLargeInt::FPLargeInt ( void )

noexcept

Default constructor.

---

FPLargeInt() [2/2]

Burger::FPLargeInt::FPLargeInt ( uint32_t uBitsNeeded )

explicitnoexcept

Constructor that sets to a default.

---

Parameters

uBitsNeeded

Number of bits of precision

See also

init(uint32_t)

Member Function Documentation

divide_return_remainder()

uint32_t Burger::FPLargeInt::divide_return_remainder ( uint32_t uDivisor )

noexcept

Divide the long number and return the remainder.

---

Given a value to perform a modulo, perform the modulo and return the remainder and leave the dividend in the data array

If uDivisor was 0, this function does nothing and returns zero

Parameters

uDivisor

32 bit unsigned number to divide the long integer

Returns

Value from 0 to uDivisor-1.

See also

multiply_return_overflow(uint32_t)

init()

void BURGER_API Burger::FPLargeInt::init ( uint32_t uBitsNeeded )

noexcept

Initialize the giant integer.

---

Reset the giant integer to defaults assuming a bit size

The maximum input value is kTotalBitsInTable -1

Parameters

uBitsNeeded

Number of bits of precision 1 through ( kTotalBitsInTable -1)

See also

FPLargeInt(uint32_t)

insert_bits_at_end()

void BURGER_API Burger::FPLargeInt::insert_bits_at_end ( uint32_t uBits, uint32_t uShiftAmount )

noexcept

Insert bits into the buffer.

---

Insert the bits from value into m_uDataTable, shifted in from the end of the array by the specified number of bits. A shift of zero or less means that none of the bits will be shifted in. A shift of one means that the high bit of value will be in the bottom of the last element of m_uDataTable; the least significant bit. A shift of kBitsPerElement means that value will be in the least significant element of m_uDataTable, and so on. It's possible a large value can be shifted out.

Parameters

uBits	32 bit data to shift in.
uShiftAmount	Number of bits to shift (0 to kTotalBitsInTable)

See also

insert_bits_at_start(uint32_t, uint32_t) or insert_chunk_bits(ChunkSize_t, uint32_t)

insert_bits_at_start()

void BURGER_API Burger::FPLargeInt::insert_bits_at_start ( uint32_t uBits, uint32_t uShiftAmount )

noexcept

Insert bits into the buffer from the top.

---

Insert the bits from value into m_uDataTable, shifted in from the beginning of the array by the specified number of bits. A shift of zero or less means that none of the bits will be shifted in. A shift of one means that the low bit of value will be in the top of the first element of m_uDataTable; the most significant bit. A shift of kBitsPerElement means that value will be in the most significant element of m_uDataTable, and so on.

Parameters

uBits	32 bit data to shift in
uShiftAmount	Number of bits to shift (0 to kTotalBitsInTable)

See also

insert_bits_at_end(uint32_t, uint32_t) or insert_chunk_bits(ChunkSize_t, uint32_t)

insert_chunk_bits()

void BURGER_API Burger::FPLargeInt::insert_chunk_bits ( ChunkSize_t uBits, uint32_t uEntryIndex )

protectednoexcept

Do the actual work of inserting bits, and updating zero flag, and extents.

---

Parameters

uBits	32 bit data to shift in
uEntryIndex	Index into the entry table

See also

InsertLowBits(uint32_t, uint32_t) or InsertTopBits(uint32_t, uint32_t)

is_not_zero()

uint_t Burger::FPLargeInt::is_not_zero ( void ) const

inlinenoexcept

Returns TRUE if the value is not zero.

---

If the giant integer's value is not zero, return TRUE

Returns

TRUE if not zero

See also

is_zero(void) const

is_zero()

uint_t Burger::FPLargeInt::is_zero ( void ) const

inlinenoexcept

Returns TRUE if the value is zero.

---

If the giant integer's value is zero, return TRUE

Returns

TRUE if zero

See also

is_not_zero(void) const

multiply_return_overflow()

uint32_t Burger::FPLargeInt::multiply_return_overflow ( uint32_t uMultiplier )

noexcept

Multiply by multiplier and return the overflow, from 0 to div-1.

---

Multiply the giant integer with a number and return the number that overflowed out.

Parameters

uMultiplier

Value to multiply the giant integer with

Returns

Overflow value, or zero if there was no overflow or the giant integer was zero.

See also

divide_return_remainder(uint32_t)

Member Data Documentation

kBitsPerElement

const uint32_t Burger::FPLargeInt::kBitsPerElement

staticconstexpr

Initial value:

=
        static_cast<uint_t>(sizeof(ChunkSize_t)) * 8

Number of bits per uint32_t.

kMaxElements

const uint32_t Burger::FPLargeInt::kMaxElements

staticconstexpr

Initial value:

=
        kTotalBitsInTable / kBitsPerElement

Number of elements in the data table.

kTotalBitsInTable

const uint32_t Burger::FPLargeInt::kTotalBitsInTable = 1024 + 64

staticconstexpr

Number of bits in the table to handle double precision.

m_bIsZero

uint32_t Burger::FPLargeInt::m_bIsZero

protected

TRUE if a zero length buffer

m_iHighestNonZeroElement

int32_t Burger::FPLargeInt::m_iHighestNonZeroElement

protected

Highest index with a non-zero entry in the data table.

m_iLowestNonZeroElement

int32_t Burger::FPLargeInt::m_iLowestNonZeroElement

protected

Lowest index with a non-zero entry in the data table.

m_uDataTable

ChunkSize_t Burger::FPLargeInt::m_uDataTable[kMaxElements]

protected

Data table of bits for locating digits.

m_uEntryCount

uint32_t Burger::FPLargeInt::m_uEntryCount

protected

Number of entries in the data table.