fml::gpuvec< T > Class Template Reference

Vector class for data held on a single GPU. More...

#include <gpuvec.hh>

Inheritance diagram for fml::gpuvec< T >:

Public Member Functions
	gpuvec (std::shared_ptr< card > gpu)
	gpuvec (std::shared_ptr< card > gpu, len_t size)
	gpuvec (std::shared_ptr< card > gpu, T *data, len_t size, bool free_on_destruct=false)
	gpuvec (const gpuvec &x)
void	resize (len_t size)
	Resize the internal object storage. More...
void	resize (std::shared_ptr< card > gpu, len_t size)
void	inherit (std::shared_ptr< card > gpu)
void	inherit (std::shared_ptr< card > gpu, T *data, len_t size, bool free_on_destruct=false)
gpuvec< T >	dupe () const
	Duplicate the object in a deep copy.
void	print (uint8_t ndigits=4, bool add_final_blank=true) const
	Copy data from a CPU object to another. More...
void	info () const
	Print some brief information about the object.
void	fill_zero ()
	Set all values to zero.
void	fill_val (const T v)
	Set all values to input value. More...
void	fill_linspace ()
	Set values to linearly spaced numbers. More...
void	fill_linspace (const T start, const T stop)
void	scale (const T s)
	Multiply all values by the input value. More...
void	pow (const T p)
	Raise every value of the vector to the given power. More...
void	rev ()
	Reverse the vector.
T	sum () const
	Sum the vector.
T	max () const
	Maximum value of the vector.
T	min () const
	Minimum value of the vector.
T	get (const len_t i) const
	Get the specified value. More...
void	set (const len_t i, const T v)
	Set the storage at the specified index with the provided value. More...
bool	operator== (const gpuvec< T > &x) const
	See if the two objects are the same. More...
bool	operator!= (const gpuvec< T > &x) const
	See if the two objects are not the same. Uses same internal logic as the == method. More...
gpuvec< T > &	operator= (const gpuvec< T > &x)
	Operator that sets the LHS to a shallow copy of the input. Desctruction of the LHS object will not result in the internal array storage being freed. More...
std::shared_ptr< card >	get_card () const
dim3	get_blockdim () const
dim3	get_griddim () const
Public Member Functions inherited from fml::univec< T >
len_t	size () const
	Number of elements in the vector.
T *	data_ptr ()
	Pointer to the internal array.
T *	data_ptr () const

Protected Attributes
std::shared_ptr< card >	c
Protected Attributes inherited from fml::univec< T >
len_t	_size
T *	data
bool	free_data

Additional Inherited Members
Protected Member Functions inherited from fml::univec< T >
bool	should_free () const
void	check_index (const len_t i) const
void	printval (const T val, uint8_t ndigits) const
void	printval (const int val, uint8_t ndigits) const

Detailed Description

template<typename T>
class fml::gpuvec< T >

Vector class for data held on a single GPU.

Template Parameters

T	should be 'int', '__half', 'float' or 'double'.

Member Function Documentation

fill_linspace()

template<typename T >

void fml::gpuvec< T >::fill_linspace

Set values to linearly spaced numbers.

Parameters

[in]

start,stop

Beginning/ending numbers. If not supplied, the vector will be filled with whole numbers from 1 to the total number of elements.

fill_val()

template<typename T >

void fml::gpuvec< T >::fill_val

(

const T

v

)

Set all values to input value.

Parameters

[in]

v

Value to set all data values to.

get()

template<typename T >

T fml::gpuvec< T >::get

(

const len_t

i

)

const

Get the specified value.

Parameters

[in]

i

The index of the desired value, 0-indexed.

Exceptions
If indices are out of bounds, the method will throw a runtime_error exception.

operator!=()

template<typename T >

bool fml::gpuvec< T >::operator!=

(

const gpuvec< T > &

x

)

const

See if the two objects are not the same. Uses same internal logic as the == method.

Parameters

[in]

Comparison

object.

operator=()

template<typename T >

fml::gpuvec< T > & fml::gpuvec< T >::operator=

(

const gpuvec< T > &

x

)

Operator that sets the LHS to a shallow copy of the input. Desctruction of the LHS object will not result in the internal array storage being freed.

Parameters

[in]

x

Setter value.

operator==()

template<typename T >

bool fml::gpuvec< T >::operator==

(

const gpuvec< T > &

x

)

const

See if the two objects are the same.

Parameters

[in]

Comparison

object.

Returns

If the sizes mismatch, then false is necessarily returned. Next, if the pointer to the internal storage arrays match, then true is necessarily returned. Otherwise the objects are compared value by value.

pow()

template<typename T >

void fml::gpuvec< T >::pow

(

const T

p

)

Raise every value of the vector to the given power.

Parameters

[in]

p

Power.

print()

template<typename REAL >

void fml::gpuvec< REAL >::print	(	uint8_t	ndigits = 4,
		bool	add_final_blank = true
	)		const

Copy data from a CPU object to another.

Parameters

[in]	ndigits	Number of decimal digits to print.
[in]	add_final_blank	Should a final blank line be printed?

resize()

template<typename T >

void fml::gpuvec< T >::resize

(

len_t

size

)

Resize the internal object storage.

Parameters

[in]

size

Length of the vector needed.

Memory Allocations
Resizing triggers a re-allocation.

Exceptions
If the reallocation fails, a bad_alloc exception will be thrown. If the input values are invalid, a runtime_error exception will be thrown.

scale()

template<typename T >

void fml::gpuvec< T >::scale

(

const T

s

)

Multiply all values by the input value.

Parameters

[in]

s

Scaling value.

set()

template<typename T >

void fml::gpuvec< T >::set	(	const len_t	i,
		const T	v
	)

Set the storage at the specified index with the provided value.

Parameters

[in]	i	The index of the desired value, 0-indexed.
[in]	v	Setter value.

Exceptions
If indices are out of bounds, the method will throw a runtime_error exception.

---

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

• fml/src/fml/gpu/gpuvec.hh