#include <tune_quda.h>

Inheritance diagram for quda::TunableVectorY:

[legend]

Collaboration diagram for quda::TunableVectorY:

Public Member Functions
	TunableVectorY (unsigned int vector_length_y)

bool	advanceBlockDim (TuneParam &param) const

void	initTuneParam (TuneParam &param) const

void	defaultTuneParam (TuneParam &param) const

void	resizeVector (int y) const

void	resizeStep (int y) const

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

virtual	~Tunable ()

virtual TuneKey	tuneKey () const =0

virtual void	apply (const cudaStream_t &stream)=0

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
virtual unsigned int	sharedBytesPerThread () const

virtual unsigned int	sharedBytesPerBlock (const TuneParam &param) const

Protected Member Functions inherited from quda::Tunable
virtual long long	flops () const =0

virtual long long	bytes () const

virtual unsigned int	minThreads () const

virtual bool	tuneGridDim () const

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 int	blockStep () const

virtual int	blockMin () const

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 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...

virtual bool	advanceSharedBytes (TuneParam &param) const

virtual bool	advanceAux (TuneParam &param) const

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

Protected Attributes
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

Detailed Description

This derived class is for algorithms that deploy a vector of computations across the y dimension of both the threads block and grid. For example this could be parity in the y dimension and checkerboarded volume in x.

Definition at line 426 of file tune_quda.h.