QUDA  v1.1.0
A library for QCD on GPUs
Public Member Functions | Protected Member Functions | List of all members
quda::TunableLocalParityReduction Class Reference

#include <tune_quda.h>

+ Inheritance diagram for quda::TunableLocalParityReduction:

Public Member Functions

bool advanceBlockDim (TuneParam &param) const
 
void initTuneParam (TuneParam &param) const
 
void defaultTuneParam (TuneParam &param) const
 
- Public Member Functions inherited from quda::Tunable
 Tunable ()
 
virtual ~Tunable ()
 
virtual TuneKey tuneKey () const =0
 
virtual void apply (const qudaStream_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

unsigned int sharedBytesPerThread () const
 
unsigned int sharedBytesPerBlock (const TuneParam &param) const
 
bool tuneGridDim () const final
 
unsigned int minGridSize () const
 
int gridStep () const
 gridStep sets the step size when iterating the grid size in advanceGridDim. More...
 
unsigned int maxBlockSize (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 tuneAuxDim () const
 
virtual bool tuneSharedBytes () const
 
virtual bool advanceGridDim (TuneParam &param) const
 
virtual unsigned int maxGridSize () const
 
virtual int blockStep () const
 
virtual int blockMin () const
 
virtual void resetBlockDim (TuneParam &param) const
 
unsigned int maxBlocksPerSM () const
 Returns the maximum number of simultaneously resident blocks per SM. We can directly query this of CUDA 11, but previously this needed to be hand coded. More...
 
unsigned int maxDynamicSharedBytesPerBlock () const
 Returns the maximum dynamic shared memory per block. 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,...)
 
bool tuned ()
 Whether the present instance has already been tuned or not. More...
 

Additional Inherited Members

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

Detailed Description

This derived class is for algorithms that deploy parity across the y dimension of the thread block with no shared memory tuning. The x threads will typically correspond to the checkboarded volume.

Definition at line 413 of file tune_quda.h.

Member Function Documentation

◆ advanceBlockDim()

bool quda::TunableLocalParityReduction::advanceBlockDim ( TuneParam param) const
inlinevirtual

Reimplemented from quda::Tunable.

Definition at line 439 of file tune_quda.h.

◆ defaultTuneParam()

void quda::TunableLocalParityReduction::defaultTuneParam ( TuneParam param) const
inlinevirtual

sets default values for when tuning is disabled

Reimplemented from quda::Tunable.

Definition at line 450 of file tune_quda.h.

◆ gridStep()

int quda::TunableLocalParityReduction::gridStep ( ) const
inlineprotectedvirtual

gridStep sets the step size when iterating the grid size in advanceGridDim.

Returns
Grid step size

Reimplemented from quda::Tunable.

Definition at line 428 of file tune_quda.h.

◆ initTuneParam()

void quda::TunableLocalParityReduction::initTuneParam ( TuneParam param) const
inlinevirtual

Reimplemented from quda::Tunable.

Definition at line 445 of file tune_quda.h.

◆ maxBlockSize()

unsigned int quda::TunableLocalParityReduction::maxBlockSize ( const TuneParam param) const
inlineprotectedvirtual

The maximum block size in the x dimension is the total number of threads divided by the size of the y dimension. Since parity is local to the thread block in the y dimension, half the max threads in the x dimension.

Reimplemented from quda::Tunable.

Definition at line 436 of file tune_quda.h.

◆ minGridSize()

unsigned int quda::TunableLocalParityReduction::minGridSize ( ) const
inlineprotectedvirtual

Reimplemented from quda::Tunable.

Definition at line 427 of file tune_quda.h.

◆ sharedBytesPerBlock()

unsigned int quda::TunableLocalParityReduction::sharedBytesPerBlock ( const TuneParam param) const
inlineprotectedvirtual

Implements quda::Tunable.

Definition at line 418 of file tune_quda.h.

◆ sharedBytesPerThread()

unsigned int quda::TunableLocalParityReduction::sharedBytesPerThread ( ) const
inlineprotectedvirtual

Implements quda::Tunable.

Definition at line 417 of file tune_quda.h.

◆ tuneGridDim()

bool quda::TunableLocalParityReduction::tuneGridDim ( ) const
inlinefinalprotectedvirtual

Reduction kernels require grid-size tuning, so enable this, and we mark as final to prevent a derived class from accidentally switching it off.

Reimplemented from quda::Tunable.

Definition at line 425 of file tune_quda.h.


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