AthCUDA::Info Class Reference

Class providing information about the CUDA devices at runtime. More...

#include <Info.h>

Collaboration diagram for AthCUDA::Info:

Public Member Functions
int	nDevices () const
	Get the number of available devices. More...
const std::vector< std::string > &	names () const
	Get the names of all of the devices. More...
const std::vector< int > &	maxThreadsPerBlock () const
	Get the maximum number of threads per execution block on all devices. More...
const std::vector< bool > &	concurrentKernels () const
	Get the multi-kernel capabilities of all devices. More...
const std::vector< std::size_t > &	totalMemory () const
	Get the total amount of memory for all devices. More...
void	print () const
	Print the properties of all detected devices. More...

Static Public Member Functions
static const Info &	instance ()
	Singleton accessor function. More...

Private Member Functions
	Info ()
	The constructor is private to implement the singleton behaviour. More...

Private Attributes
int	m_nDevices = 0
	The number of CUDA devices available at runtime. More...
std::vector< std::string >	m_names
	The names of all available devices. More...
std::vector< int >	m_maxThreadsPerBlock
	The maximum number of threads per block for each available device. More...
std::vector< bool >	m_concurrentKernels
	The multi-kernel capabilities of all devices. More...
std::vector< std::size_t >	m_totalMemory
	The total amount of memory for all devices. More...

Detailed Description

Class providing information about the CUDA devices at runtime.

This is meant to be most useful when we build the code with CUDA available, but then try to execute the code on a machine that has no GPUs.

Author

Attila Krasznahorkay Attil.nosp@m.a.Kr.nosp@m.aszna.nosp@m.hork.nosp@m.ay@ce.nosp@m.rn.c.nosp@m.h

Definition at line 24 of file Info.h.

Constructor & Destructor Documentation

Info()

AthCUDA::Info::Info ( )

private

The constructor is private to implement the singleton behaviour.

Member Function Documentation

concurrentKernels()

const std::vector< bool >& AthCUDA::Info::concurrentKernels

(

)

const

Get the multi-kernel capabilities of all devices.

instance()

static const Info& AthCUDA::Info::instance ( )

static

Singleton accessor function.

maxThreadsPerBlock()

const std::vector< int >& AthCUDA::Info::maxThreadsPerBlock

(

)

const

Get the maximum number of threads per execution block on all devices.

names()

const std::vector< std::string >& AthCUDA::Info::names

(

)

const

Get the names of all of the devices.

nDevices()

int AthCUDA::Info::nDevices

(

)

const

Get the number of available devices.

print()

void AthCUDA::Info::print

(

)

const

Print the properties of all detected devices.

totalMemory()

const std::vector< std::size_t >& AthCUDA::Info::totalMemory

(

)

const

Get the total amount of memory for all devices.

Member Data Documentation

m_concurrentKernels

std::vector< bool > AthCUDA::Info::m_concurrentKernels

private

The multi-kernel capabilities of all devices.

Definition at line 56 of file Info.h.

m_maxThreadsPerBlock

std::vector< int > AthCUDA::Info::m_maxThreadsPerBlock

private

The maximum number of threads per block for each available device.

Definition at line 54 of file Info.h.

m_names

std::vector< std::string > AthCUDA::Info::m_names

private

The names of all available devices.

Definition at line 52 of file Info.h.

m_nDevices

int AthCUDA::Info::m_nDevices = 0

private

The number of CUDA devices available at runtime.

Definition at line 50 of file Info.h.

m_totalMemory

std::vector< std::size_t > AthCUDA::Info::m_totalMemory

private

The total amount of memory for all devices.

Definition at line 58 of file Info.h.

---

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

• Info.h