Inheritance diagram for quda::Dslash5< Float, nColor, Arg >:

Collaboration diagram for quda::Dslash5< Float, nColor, Arg >:

[legend]

Public Member Functions
	Dslash5 (Arg &arg, const ColorSpinorField &meta)

virtual	~Dslash5 ()

template<typename T >
void	launch (T *f, const TuneParam &tp, Arg &arg, const cudaStream_t &stream)

void	apply (const cudaStream_t &stream)

void	initTuneParam (TuneParam &param) const

void	defaultTuneParam (TuneParam &param) const

TuneKey	tuneKey () const

Public Member Functions inherited from quda::TunableVectorYZ
	TunableVectorYZ (unsigned int vector_length_y, unsigned int vector_length_z)

bool	advanceBlockDim (TuneParam &param) const

void	resizeVector (int y, int z) const

void	resizeStep (int y, int z) const

Public Member Functions inherited from quda::TunableVectorY
	TunableVectorY (unsigned int vector_length_y)

void	resizeVector (int y) const

void	resizeStep (int y) const

Public Member Functions inherited from quda::Tunable
	Tunable ()

virtual	~Tunable ()

virtual void	preTune ()

virtual void	postTune ()

virtual int	tuningIter () const

virtual std::string	paramString (const TuneParam &param) const

virtual std::string	perfString (float time) const

virtual bool	advanceTuneParam (TuneParam &param) const

void	checkLaunchParam (TuneParam &param)

CUresult	jitifyError () const

CUresult &	jitifyError ()

Protected Member Functions
long long	flops () const

long long	bytes () const

bool	tuneGridDim () const

unsigned int	minThreads () const

int	blockStep () const

int	blockMin () const

unsigned int	sharedBytesPerThread () const

unsigned int	maxSharedBytesPerBlock () const
	The maximum shared memory that a CUDA thread block can use in the autotuner. This isn't necessarily the same as maxDynamicSharedMemoryPerBlock since that may need explicit opt in to enable (by calling setMaxDynamicSharedBytes for the kernel in question). If the CUDA kernel in question does this opt in then this function can be overloaded to return maxDynamicSharedBytesPerBlock. More...

Protected Member Functions inherited from quda::TunableVectorY
virtual unsigned int	sharedBytesPerBlock (const TuneParam &param) const

Protected Member Functions inherited from quda::Tunable
virtual bool	tuneAuxDim () const

virtual bool	tuneSharedBytes () const

virtual bool	advanceGridDim (TuneParam &param) const

virtual unsigned int	maxBlockSize (const TuneParam &param) const

virtual unsigned int	maxGridSize () const

virtual unsigned int	minGridSize () const

virtual int	gridStep () const
	gridStep sets the step size when iterating the grid size in advanceGridDim. More...

virtual void	resetBlockDim (TuneParam &param) const

unsigned int	maxBlocksPerSM () const
	For some reason this can't be queried from the device properties, so here we set set this. Based on Table 14 of the CUDA Programming Guide 10.0 (Technical Specifications per Compute Capability) More...

template<typename F >
void	setMaxDynamicSharedBytesPerBlock (F *func) const
	Enable the maximum dynamic shared bytes for the kernel "func" (values given by maxDynamicSharedBytesPerBlock()). More...

unsigned int	maxDynamicSharedBytesPerBlock () const
	This can't be correctly queried in CUDA for all architectures so here we set set this. Based on Table 14 of the CUDA Programming Guide 10.0 (Technical Specifications per Compute Capability). More...

virtual bool	advanceSharedBytes (TuneParam &param) const

virtual bool	advanceAux (TuneParam &param) const

int	writeAuxString (const char *format,...)

Protected Attributes
Arg &	arg

const ColorSpinorField &	meta

Protected Attributes inherited from quda::TunableVectorY
unsigned int	vector_length_y

unsigned int	step_y

bool	tune_block_x

Protected Attributes inherited from quda::Tunable
char	aux [TuneKey::aux_n]

CUresult	jitify_error

Static Protected Attributes
static constexpr bool	shared = true

static constexpr bool	var_inverse = true

Detailed Description

template<typename Float, int nColor, typename Arg>
class quda::Dslash5< Float, nColor, Arg >

Definition at line 20 of file dslash5_domain_wall.cu.

Constructor & Destructor Documentation

◆ Dslash5()

template<typename Float, int nColor, typename Arg>

quda::Dslash5< Float, nColor, Arg >::Dslash5	(	Arg &	arg,
		const ColorSpinorField &	meta
	)

inline

Definition at line 103 of file dslash5_domain_wall.cu.

References quda::Tunable::aux, quda::LatticeField::AuxString(), quda::DSLASH5_DWF, quda::DSLASH5_MOBIUS, quda::DSLASH5_MOBIUS_PRE, errorQuda, quda::M5_INV_DWF, quda::M5_INV_MOBIUS, and quda::M5_INV_ZMOBIUS.

Here is the call graph for this function:

◆ ~Dslash5()

template<typename Float, int nColor, typename Arg>

virtual quda::Dslash5< Float, nColor, Arg >::~Dslash5 ( )

inlinevirtual

Definition at line 118 of file dslash5_domain_wall.cu.

Member Function Documentation

◆ apply()

template<typename Float, int nColor, typename Arg>

void quda::Dslash5< Float, nColor, Arg >::apply ( const cudaStream_t & stream )

inlinevirtual

Implements quda::Tunable.

Definition at line 130 of file dslash5_domain_wall.cu.

References quda::DSLASH5_DWF, quda::DSLASH5_MOBIUS, quda::DSLASH5_MOBIUS_PRE, errorQuda, getTuning(), getVerbosity(), quda::Dslash5< Float, nColor, Arg >::launch(), quda::LatticeField::Location(), quda::M5_INV_DWF, quda::M5_INV_MOBIUS, quda::M5_INV_ZMOBIUS, QUDA_CPU_FIELD_LOCATION, and quda::tuneLaunch().

Referenced by quda::ApplyDslash5().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ blockMin()

template<typename Float, int nColor, typename Arg>

int quda::Dslash5< Float, nColor, Arg >::blockMin ( ) const

inlineprotectedvirtual

Reimplemented from quda::Tunable.

Definition at line 80 of file dslash5_domain_wall.cu.

◆ blockStep()

template<typename Float, int nColor, typename Arg>

int quda::Dslash5< Float, nColor, Arg >::blockStep ( ) const

inlineprotectedvirtual

Reimplemented from quda::Tunable.

Definition at line 79 of file dslash5_domain_wall.cu.

◆ bytes()

template<typename Float, int nColor, typename Arg>

long long quda::Dslash5< Float, nColor, Arg >::bytes ( ) const

inlineprotectedvirtual

Reimplemented from quda::Tunable.

Definition at line 62 of file dslash5_domain_wall.cu.

References quda::DSLASH5_DWF, quda::DSLASH5_MOBIUS, quda::DSLASH5_MOBIUS_PRE, errorQuda, Ls, quda::M5_INV_DWF, quda::M5_INV_MOBIUS, quda::M5_INV_ZMOBIUS, and quda::ColorSpinorField::X().

Here is the call graph for this function:

◆ defaultTuneParam()

template<typename Float, int nColor, typename Arg>

void quda::Dslash5< Float, nColor, Arg >::defaultTuneParam ( TuneParam & param ) const

inlinevirtual

sets default values for when tuning is disabled

Reimplemented from quda::TunableVectorYZ.

Definition at line 202 of file dslash5_domain_wall.cu.

References quda::TuneParam::block, quda::TunableVectorYZ::defaultTuneParam(), quda::TuneParam::grid, quda::M5_INV_DWF, quda::M5_INV_MOBIUS, quda::M5_INV_ZMOBIUS, quda::TuneParam::shared_bytes, and quda::Dslash5< Float, nColor, Arg >::sharedBytesPerThread().

Here is the call graph for this function:

◆ flops()

template<typename Float, int nColor, typename Arg>

long long quda::Dslash5< Float, nColor, Arg >::flops ( ) const

inlineprotectedvirtual

Implements quda::Tunable.

Definition at line 31 of file dslash5_domain_wall.cu.

References quda::DSLASH5_DWF, quda::DSLASH5_MOBIUS, quda::DSLASH5_MOBIUS_PRE, errorQuda, Ls, quda::M5_INV_DWF, quda::M5_INV_MOBIUS, quda::M5_INV_ZMOBIUS, quda::ColorSpinorField::Ncolor(), quda::ColorSpinorField::Nspin(), quda::ColorSpinorField::Volume(), and quda::ColorSpinorField::X().

Here is the call graph for this function:

◆ initTuneParam()

template<typename Float, int nColor, typename Arg>

void quda::Dslash5< Float, nColor, Arg >::initTuneParam ( TuneParam & param ) const

inlinevirtual

Reimplemented from quda::TunableVectorYZ.

Definition at line 192 of file dslash5_domain_wall.cu.

References quda::TuneParam::block, quda::TuneParam::grid, quda::TunableVectorYZ::initTuneParam(), quda::M5_INV_DWF, quda::M5_INV_MOBIUS, quda::M5_INV_ZMOBIUS, quda::TuneParam::shared_bytes, and quda::Dslash5< Float, nColor, Arg >::sharedBytesPerThread().

Here is the call graph for this function:

◆ launch()

template<typename Float, int nColor, typename Arg>

template<typename T >

void quda::Dslash5< Float, nColor, Arg >::launch	(	T *	f,
		const TuneParam &	tp,
		Arg &	arg,
		const cudaStream_t &	stream
	)

inline

Definition at line 120 of file dslash5_domain_wall.cu.

References quda::TuneParam::block, quda::TuneParam::grid, quda::M5_INV_DWF, quda::M5_INV_MOBIUS, quda::M5_INV_ZMOBIUS, quda::qudaLaunchKernel(), quda::Tunable::setMaxDynamicSharedBytesPerBlock(), and quda::TuneParam::shared_bytes.

Referenced by quda::Dslash5< Float, nColor, Arg >::apply().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ maxSharedBytesPerBlock()

template<typename Float, int nColor, typename Arg>

unsigned int quda::Dslash5< Float, nColor, Arg >::maxSharedBytesPerBlock ( ) const

inlineprotectedvirtual

The maximum shared memory that a CUDA thread block can use in the autotuner. This isn't necessarily the same as maxDynamicSharedMemoryPerBlock since that may need explicit opt in to enable (by calling setMaxDynamicSharedBytes for the kernel in question). If the CUDA kernel in question does this opt in then this function can be overloaded to return maxDynamicSharedBytesPerBlock.