QUDA
1.0.0
|
Public Member Functions | |
MultiReduceCuda (doubleN result[], SpinorX X[], SpinorY Y[], SpinorZ Z[], SpinorW W[], Reducer &r, std::vector< ColorSpinorField *> &x, std::vector< ColorSpinorField *> &y, std::vector< ColorSpinorField *> &z, std::vector< ColorSpinorField *> &w, int NYW, int length) | |
TuneKey | tuneKey () const |
void | apply (const cudaStream_t &stream) |
bool | advanceGridDim (TuneParam ¶m) const |
void | initTuneParam (TuneParam ¶m) const |
void | defaultTuneParam (TuneParam ¶m) const |
void | preTune () |
void | postTune () |
long long | flops () const |
long long | bytes () const |
int | tuningIter () const |
![]() | |
Tunable () | |
virtual | ~Tunable () |
virtual std::string | paramString (const TuneParam ¶m) const |
virtual std::string | perfString (float time) const |
virtual bool | advanceTuneParam (TuneParam ¶m) const |
void | checkLaunchParam (TuneParam ¶m) |
CUresult | jitifyError () const |
CUresult & | jitifyError () |
Private Member Functions | |
unsigned int | sharedBytesPerThread () const |
unsigned int | sharedBytesPerBlock (const TuneParam ¶m) const |
virtual bool | advanceSharedBytes (TuneParam ¶m) const |
unsigned int | maxBlockSize (const TuneParam ¶m) const |
Private Attributes | |
const int | NYW |
int | nParity |
MultiReduceArg< NXZ, ReduceType, SpinorX, SpinorY, SpinorZ, SpinorW, Reducer > | arg |
doubleN * | result |
std::vector< ColorSpinorField * > & | x |
std::vector< ColorSpinorField * > & | y |
std::vector< ColorSpinorField * > & | z |
std::vector< ColorSpinorField * > & | w |
char * | Y_h [MAX_MULTI_BLAS_N] |
char * | W_h [MAX_MULTI_BLAS_N] |
char * | Ynorm_h [MAX_MULTI_BLAS_N] |
char * | Wnorm_h [MAX_MULTI_BLAS_N] |
Additional Inherited Members | |
![]() | |
virtual unsigned int | minThreads () const |
virtual bool | tuneGridDim () const |
virtual bool | tuneAuxDim () const |
virtual bool | tuneSharedBytes () 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 ¶m) const |
virtual bool | advanceBlockDim (TuneParam ¶m) 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 | advanceAux (TuneParam ¶m) const |
int | writeAuxString (const char *format,...) |
![]() | |
char | aux [TuneKey::aux_n] |
CUresult | jitify_error |
Definition at line 111 of file multi_reduce_quda.cu.
|
inline |
Definition at line 144 of file multi_reduce_quda.cu.
References quda::blas::getFastReduce(), and V.
|
inlinevirtual |
Reimplemented from quda::Tunable.
Definition at line 220 of file multi_reduce_quda.cu.
References quda::Tunable::advanceGridDim(), deviceProp, and errorQuda.
|
inlineprivatevirtual |
The goal here is to throttle the number of thread blocks per SM by over-allocating shared memory (in order to improve L2 utilization, etc.). We thus request the smallest amount of dynamic shared memory that guarantees throttling to a given number of blocks, in order to allow some extra leeway.
Reimplemented from quda::Tunable.
Definition at line 128 of file multi_reduce_quda.cu.
References quda::TuneParam::block, and quda::TuneParam::shared_bytes.
|
inlinevirtual |
Implements quda::Tunable.
Definition at line 191 of file multi_reduce_quda.cu.
References quda::arg(), quda::TuneParam::block, getTuning(), getVerbosity(), quda::TuneParam::grid, param, quda::stream, and quda::tuneLaunch().
Referenced by quda::blas::multiReduce().
|
inlinevirtual |
Reimplemented from quda::Tunable.
Definition at line 264 of file multi_reduce_quda.cu.
References quda::blas::MultiReduceArg< NXZ, ReduceType, SpinorX, SpinorY, SpinorZ, SpinorW, Reducer >::r.
Referenced by quda::blas::multiReduce().
|
inlinevirtual |
sets default values for when tuning is disabled
Reimplemented from quda::Tunable.
Definition at line 235 of file multi_reduce_quda.cu.
References quda::TuneParam::block, quda::Tunable::defaultTuneParam(), and quda::TuneParam::grid.
|
inlinevirtual |
Implements quda::Tunable.
Definition at line 259 of file multi_reduce_quda.cu.
References quda::blas::MultiReduceArg< NXZ, ReduceType, SpinorX, SpinorY, SpinorZ, SpinorW, Reducer >::length, and quda::blas::MultiReduceArg< NXZ, ReduceType, SpinorX, SpinorY, SpinorZ, SpinorW, Reducer >::r.
Referenced by quda::blas::multiReduce().
|
inlinevirtual |
Reimplemented from quda::Tunable.
Definition at line 227 of file multi_reduce_quda.cu.
References quda::TuneParam::block, quda::TuneParam::grid, and quda::Tunable::initTuneParam().
|
inlineprivatevirtual |
Reimplemented from quda::Tunable.
Definition at line 141 of file multi_reduce_quda.cu.
References deviceProp.
|
inlinevirtual |
Reimplemented from quda::Tunable.
Definition at line 251 of file multi_reduce_quda.cu.
References quda::blas::MultiReduceArg< NXZ, ReduceType, SpinorX, SpinorY, SpinorZ, SpinorW, Reducer >::W, and quda::blas::MultiReduceArg< NXZ, ReduceType, SpinorX, SpinorY, SpinorZ, SpinorW, Reducer >::Y.
|
inlinevirtual |
Reimplemented from quda::Tunable.
Definition at line 243 of file multi_reduce_quda.cu.
References quda::blas::MultiReduceArg< NXZ, ReduceType, SpinorX, SpinorY, SpinorZ, SpinorW, Reducer >::W, and quda::blas::MultiReduceArg< NXZ, ReduceType, SpinorX, SpinorY, SpinorZ, SpinorW, Reducer >::Y.
|
inlineprivatevirtual |
Implements quda::Tunable.
Definition at line 126 of file multi_reduce_quda.cu.
|
inlineprivatevirtual |
Implements quda::Tunable.
Definition at line 125 of file multi_reduce_quda.cu.
|
inlinevirtual |
Implements quda::Tunable.
Definition at line 182 of file multi_reduce_quda.cu.
References quda::TuneKey::name_n, and quda::blas::MultiReduceArg< NXZ, ReduceType, SpinorX, SpinorY, SpinorZ, SpinorW, Reducer >::r.
|
inlinevirtual |
Reimplemented from quda::Tunable.
Definition at line 270 of file multi_reduce_quda.cu.
|
private |
Definition at line 117 of file multi_reduce_quda.cu.
|
private |
Definition at line 116 of file multi_reduce_quda.cu.
|
private |
Definition at line 115 of file multi_reduce_quda.cu.
|
private |
Definition at line 118 of file multi_reduce_quda.cu.
|
private |
Definition at line 120 of file multi_reduce_quda.cu.
|
private |
Definition at line 123 of file multi_reduce_quda.cu.
|
private |
Definition at line 123 of file multi_reduce_quda.cu.
|
private |
Definition at line 120 of file multi_reduce_quda.cu.
|
private |
Definition at line 120 of file multi_reduce_quda.cu.
|
private |
Definition at line 123 of file multi_reduce_quda.cu.
|
private |
Definition at line 123 of file multi_reduce_quda.cu.
|
private |
Definition at line 120 of file multi_reduce_quda.cu.