QUDA
1.0.0
|
Public Member Functions | |
SharedDslashCuda (cudaColorSpinorField *out, const cudaColorSpinorField *in, const cudaColorSpinorField *x, const GaugeField &gauge, int parity, int dagger, const int *commOverride) | |
virtual | ~SharedDslashCuda () |
![]() | |
DslashCuda (cudaColorSpinorField *out, const cudaColorSpinorField *in, const cudaColorSpinorField *x, const GaugeField &gauge, const int parity, const int dagger, const int *commOverride) | |
virtual | ~DslashCuda () |
virtual TuneKey | tuneKey () const |
const char * | getAux (KernelType type) const |
void | setAux (KernelType type, const char *aux_) |
void | augmentAux (KernelType type, const char *extra) |
virtual int | Nface () const |
int | Dagger () const |
virtual void | preTune () |
virtual void | postTune () |
virtual long long | flops () const |
virtual long long | bytes () const |
Additional Inherited Members | |
![]() | |
DslashParam | dslashParam |
![]() | |
unsigned int | sharedBytesPerBlock (const TuneParam ¶m) const |
bool | tuneGridDim () const |
bool | tuneAuxDim () const |
unsigned int | minThreads () const |
void | fillAuxBase () |
Set the base strings used by the different dslash kernel types for autotuning. More... | |
void | fillAux (KernelType kernel_type, const char *kernel_str) |
Specialize the auxiliary strings for each kernel type. More... | |
void | setParam () |
Set the dslashParam for the current multi-GPU parameters (set these at the last minute to ensure we always use the correct ones while policy autotuning). More... | |
![]() | |
cudaColorSpinorField * | out |
const cudaColorSpinorField * | in |
const cudaColorSpinorField * | x |
const GaugeField & | gauge |
const QudaReconstructType | reconstruct |
char * | saveOut |
char * | saveOutNorm |
const int | dagger |
char | aux_base [TuneKey::aux_n] |
char | aux [8][TuneKey::aux_n] |
![]() | |
static bool | init = false |
static char | ghost_str [TuneKey::aux_n] |
This derived class is specifically for driving the Dslash kernels that use shared memory blocking. This only applies on Fermi and upwards, and only for the interior kernels. For pre-Fermi architectures
Definition at line 859 of file dslash_quda.cuh.
|
inline |
Definition at line 861 of file dslash_quda.cuh.
|
inlinevirtual |
Definition at line 865 of file dslash_quda.cuh.