QUDA  1.0.0
•All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Macros
Public Member Functions | Private Types | Private Member Functions | Private Attributes | List of all members
quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal > Class Template Reference
Inheritance diagram for quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >:
Inheritance graph
[legend]
Collaboration diagram for quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >:
Collaboration graph
[legend]

Public Member Functions

 TileSizeTune (Complex *result, vec &x, vec &y, vec &z, vec &w, bool hermitian, bool Anorm=false)
 
virtual ~TileSizeTune ()
 
void apply (const cudaStream_t &stream)
 
bool advanceAux (TuneParam &param) const
 
bool advanceTuneParam (TuneParam &param) const
 
void initTuneParam (TuneParam &param) const
 
void defaultTuneParam (TuneParam &param) const
 
TuneKey tuneKey () const
 
long long flops () const
 
long long bytes () const
 
void preTune ()
 
void postTune ()
 
- Public Member Functions inherited from quda::Tunable
 Tunable ()
 
virtual ~Tunable ()
 
virtual int tuningIter () const
 
virtual std::string paramString (const TuneParam &param) const
 
virtual std::string perfString (float time) const
 
void checkLaunchParam (TuneParam &param)
 
CUresult jitifyError () const
 
CUresult & jitifyError ()
 

Private Types

typedef std::vector< ColorSpinorField * > vec
 

Private Member Functions

unsigned int sharedBytesPerThread () const
 
unsigned int sharedBytesPerBlock (const TuneParam &param) const
 

Private Attributes

Complex * result
 
vec & x
 
vec & y
 
vec & z
 
vec & w
 
bool hermitian
 
bool Anorm
 
unsigned int max_tile_size
 

Additional Inherited Members

- Protected Member Functions inherited from quda::Tunable
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
 
virtual bool advanceBlockDim (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
 
int writeAuxString (const char *format,...)
 
- Protected Attributes inherited from quda::Tunable
char aux [TuneKey::aux_n]
 
CUresult jitify_error
 

Detailed Description

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
class quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >

Definition at line 880 of file multi_reduce_quda.cu.

Member Typedef Documentation

◆ vec

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
typedef std::vector<ColorSpinorField*> quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::vec
private

Definition at line 881 of file multi_reduce_quda.cu.

Constructor & Destructor Documentation

◆ TileSizeTune()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::TileSizeTune ( Complex *  result,
vec &  x,
vec &  y,
vec &  z,
vec &  w,
bool  hermitian,
bool  Anorm = false 
)
inline

Definition at line 893 of file multi_reduce_quda.cu.

References quda::disableProfileCount(), quda::enableProfileCount(), quda::blas::end(), quda::getTuneCache(), getTuning(), MAX_MULTI_BLAS_N, quda::setPolicyTuning(), quda::size, and quda::u64toa().

Here is the call graph for this function:

◆ ~TileSizeTune()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
virtual quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::~TileSizeTune ( )
inlinevirtual

Definition at line 961 of file multi_reduce_quda.cu.

References quda::setPolicyTuning().

Here is the call graph for this function:

Member Function Documentation

◆ advanceAux()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
bool quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::advanceAux ( TuneParam &  param) const
inlinevirtual

Reimplemented from quda::Tunable.

Definition at line 975 of file multi_reduce_quda.cu.

References quda::TuneParam::aux, and MAX_MULTI_BLAS_N.

◆ advanceTuneParam()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
bool quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::advanceTuneParam ( TuneParam &  param) const
inlinevirtual

Reimplemented from quda::Tunable.

Definition at line 1007 of file multi_reduce_quda.cu.

◆ apply()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
void quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::apply ( const cudaStream_t &  stream)
inlinevirtual

Implements quda::Tunable.

Definition at line 963 of file multi_reduce_quda.cu.

References quda::TuneParam::aux, getTuning(), getVerbosity(), and quda::tuneLaunch().

Referenced by quda::blas::cDotProduct(), quda::blas::cDotProductCopy(), quda::blas::hDotProduct(), and quda::blas::hDotProduct_Anorm().

Here is the call graph for this function:
Here is the caller graph for this function:

◆ bytes()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
long long quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::bytes ( ) const
inlinevirtual

Reimplemented from quda::Tunable.

Definition at line 1025 of file multi_reduce_quda.cu.

◆ defaultTuneParam()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
void quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::defaultTuneParam ( TuneParam &  param) const
inlinevirtual

sets default values for when tuning is disabled

Reimplemented from quda::Tunable.

Definition at line 1014 of file multi_reduce_quda.cu.

References quda::TuneParam::aux, and quda::Tunable::defaultTuneParam().

Here is the call graph for this function:

◆ flops()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
long long quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::flops ( ) const
inlinevirtual

Implements quda::Tunable.

Definition at line 1024 of file multi_reduce_quda.cu.

◆ initTuneParam()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
void quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::initTuneParam ( TuneParam &  param) const
inlinevirtual

Reimplemented from quda::Tunable.

Definition at line 1009 of file multi_reduce_quda.cu.

References quda::TuneParam::aux, and quda::Tunable::initTuneParam().

Here is the call graph for this function:

◆ postTune()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
void quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::postTune ( )
inlinevirtual

Reimplemented from quda::Tunable.

Definition at line 1028 of file multi_reduce_quda.cu.

◆ preTune()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
void quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::preTune ( )
inlinevirtual

Reimplemented from quda::Tunable.

Definition at line 1027 of file multi_reduce_quda.cu.

◆ sharedBytesPerBlock()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
unsigned int quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::sharedBytesPerBlock ( const TuneParam &  param) const
inlineprivatevirtual

Implements quda::Tunable.

Definition at line 888 of file multi_reduce_quda.cu.

◆ sharedBytesPerThread()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
unsigned int quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::sharedBytesPerThread ( ) const
inlineprivatevirtual

Implements quda::Tunable.

Definition at line 887 of file multi_reduce_quda.cu.

◆ tuneKey()

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
TuneKey quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::tuneKey ( ) const
inlinevirtual

Implements quda::Tunable.

Definition at line 1020 of file multi_reduce_quda.cu.

Member Data Documentation

◆ Anorm

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
bool quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::Anorm
private

Definition at line 885 of file multi_reduce_quda.cu.

◆ hermitian

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
bool quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::hermitian
private

Definition at line 884 of file multi_reduce_quda.cu.

◆ max_tile_size

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
unsigned int quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::max_tile_size
private

Definition at line 890 of file multi_reduce_quda.cu.

◆ result

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
Complex* quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::result
private

Definition at line 882 of file multi_reduce_quda.cu.

◆ w

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
vec & quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::w
private

Definition at line 883 of file multi_reduce_quda.cu.

◆ x

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
vec& quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::x
private

Definition at line 883 of file multi_reduce_quda.cu.

◆ y

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
vec & quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::y
private

Definition at line 883 of file multi_reduce_quda.cu.

◆ z

template<template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerDiagonal, typename writeDiagonal, template< int MXZ, typename ReducerType, typename Float, typename FloatN > class ReducerOffDiagonal, typename writeOffDiagonal>
vec & quda::blas::TileSizeTune< ReducerDiagonal, writeDiagonal, ReducerOffDiagonal, writeOffDiagonal >::z
private

Definition at line 883 of file multi_reduce_quda.cu.

The documentation for this class was generated from the following file:
Generated by   doxygen 1.8.13