QUDA
1.0.0
|
Public Member Functions | |
DslashCoarsePolicyTune (DslashCoarseLaunch &dslash) | |
virtual | ~DslashCoarsePolicyTune () |
void | apply (const cudaStream_t &stream) |
int | tuningIter () const |
bool | advanceAux (TuneParam ¶m) const |
bool | advanceTuneParam (TuneParam ¶m) const |
void | initTuneParam (TuneParam ¶m) const |
void | defaultTuneParam (TuneParam ¶m) const |
TuneKey | tuneKey () const |
long long | flops () const |
long long | bytes () const |
![]() | |
Tunable () | |
virtual | ~Tunable () |
virtual void | preTune () |
virtual void | postTune () |
virtual std::string | paramString (const TuneParam ¶m) const |
virtual std::string | perfString (float time) const |
void | checkLaunchParam (TuneParam ¶m) |
CUresult | jitifyError () const |
CUresult & | jitifyError () |
Private Member Functions | |
bool | tuneGridDim () const |
bool | tuneAuxDim () const |
unsigned int | sharedBytesPerThread () const |
unsigned int | sharedBytesPerBlock (const TuneParam ¶m) const |
Private Attributes | |
DslashCoarseLaunch & | dslash |
Additional Inherited Members | |
![]() | |
virtual unsigned int | minThreads () const |
virtual bool | tuneSharedBytes () const |
virtual bool | advanceGridDim (TuneParam ¶m) const |
virtual unsigned int | maxBlockSize (const TuneParam ¶m) 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 | advanceSharedBytes (TuneParam ¶m) const |
int | writeAuxString (const char *format,...) |
![]() | |
char | aux [TuneKey::aux_n] |
CUresult | jitify_error |
Definition at line 610 of file dslash_coarse.cu.
|
inline |
Definition at line 620 of file dslash_coarse.cu.
References quda::LatticeField::AuxString(), quda::DslashCoarseLaunch::clover, comm_config_string(), comm_dim_partitioned_string(), comm_dim_topology_string(), comm_gdr_enabled(), quda::DslashCoarseLaunch::commDim, quda::disableProfileCount(), quda::DslashCoarseLaunch::dslash, quda::DSLASH_COARSE_BASIC, quda::DSLASH_COARSE_GDR, quda::DSLASH_COARSE_GDR_RECV, quda::DSLASH_COARSE_GDR_SEND, quda::DSLASH_COARSE_GDR_SEND_ZERO_COPY_READ, quda::DSLASH_COARSE_POLICY_DISABLED, quda::DSLASH_COARSE_ZERO_COPY, quda::DSLASH_COARSE_ZERO_COPY_PACK, quda::DSLASH_COARSE_ZERO_COPY_PACK_GDR_RECV, quda::DSLASH_COARSE_ZERO_COPY_READ, quda::enable_policy(), quda::enableProfileCount(), quda::blas::end(), errorQuda, quda::first_active_policy, quda::getTuneCache(), getTuning(), quda::DslashCoarseLaunch::halo_precision, quda::i32toa(), quda::DslashCoarseLaunch::inA, quda::policies(), prec_str, quda::LatticeField::Precision(), quda::setPolicyTuning(), and quda::DslashCoarseLaunch::Y.
|
inlinevirtual |
Definition at line 703 of file dslash_coarse.cu.
References quda::setPolicyTuning().
|
inlinevirtual |
Reimplemented from quda::Tunable.
Definition at line 715 of file dslash_coarse.cu.
References quda::TuneParam::aux, quda::DSLASH_COARSE_POLICY_DISABLED, and quda::policies().
|
inlinevirtual |
Reimplemented from quda::Tunable.
Definition at line 725 of file dslash_coarse.cu.
|
inlinevirtual |
Implements quda::Tunable.
Definition at line 705 of file dslash_coarse.cu.
References quda::TuneParam::aux, quda::DSLASH_COARSE_POLICY_DISABLED, errorQuda, getTuning(), getVerbosity(), quda::policies(), and quda::tuneLaunch().
Referenced by quda::ApplyCoarse().
|
inlinevirtual |
Reimplemented from quda::Tunable.
Definition at line 756 of file dslash_coarse.cu.
References quda::GaugeField::Bytes(), quda::ColorSpinorField::Bytes(), quda::DslashCoarseLaunch::clover, quda::DslashCoarseLaunch::dslash, quda::DslashCoarseLaunch::inA, quda::DslashCoarseLaunch::inB, quda::DslashCoarseLaunch::out, quda::ColorSpinorField::SiteSubset(), quda::LatticeField::Stride(), quda::LatticeField::VolumeCB(), quda::DslashCoarseLaunch::X, and quda::DslashCoarseLaunch::Y.
|
inlinevirtual |
sets default values for when tuning is disabled
Reimplemented from quda::Tunable.
Definition at line 735 of file dslash_coarse.cu.
References quda::TuneParam::aux, quda::Tunable::defaultTuneParam(), and quda::first_active_policy.
|
inlinevirtual |
Implements quda::Tunable.
Definition at line 747 of file dslash_coarse.cu.
References quda::DslashCoarseLaunch::clover, quda::DslashCoarseLaunch::dslash, quda::DslashCoarseLaunch::inA, quda::ColorSpinorField::Ncolor(), quda::ColorSpinorField::Nspin(), quda::ColorSpinorField::SiteSubset(), volumeCB, and quda::ColorSpinorField::VolumeCB().
|
inlinevirtual |
Reimplemented from quda::Tunable.
Definition at line 727 of file dslash_coarse.cu.
References quda::TuneParam::aux, quda::first_active_policy, and quda::Tunable::initTuneParam().
|
inlineprivatevirtual |
Implements quda::Tunable.
Definition at line 617 of file dslash_coarse.cu.
|
inlineprivatevirtual |
Implements quda::Tunable.
Definition at line 616 of file dslash_coarse.cu.
|
inlineprivatevirtual |
Reimplemented from quda::Tunable.
Definition at line 615 of file dslash_coarse.cu.
|
inlineprivatevirtual |
Reimplemented from quda::Tunable.
Definition at line 614 of file dslash_coarse.cu.
|
inlinevirtual |
Implements quda::Tunable.
Definition at line 743 of file dslash_coarse.cu.
References quda::DslashCoarseLaunch::inA, and quda::LatticeField::VolString().
|
inlinevirtual |
Reimplemented from quda::Tunable.
Definition at line 713 of file dslash_coarse.cu.
|
private |
Definition at line 612 of file dslash_coarse.cu.