The interface which should be implemented to provide optimised computation functions for implementations of RooAbsReal::evaluateSpan().
The class RooBatchComputeInterface provides the mechanism for external modules (like RooFit) to call functions from the library. The power lies in the virtual functions that can resolve to different implementations for the functionality; for example, calling a function through dispatchCuda will resolve to efficient cuda implementations.
This interface contains the signatures of the compute functions of every PDF that has an optimised implementation available. These are the functions that perform the actual computations in batches.
Several implementations of this interface may be provided, e.g. SSE, AVX, AVX2 etc. At run time, the fastest implementation of this interface is selected, and using a virtual call, the computation is dispatched to the best backend.
Definition at line 61 of file RooBatchCompute.h.
Public Member Functions | |
virtual | ~RooBatchComputeInterface ()=default |
virtual Architecture | architecture () const =0 |
virtual std::string | architectureName () const =0 |
virtual void | compute (cudaStream_t *, Computer, RestrictArr, size_t, const VarVector &, const ArgVector &={})=0 |
virtual float | cudaEventElapsedTime (cudaEvent_t *, cudaEvent_t *) |
virtual void | cudaEventRecord (cudaEvent_t *, cudaStream_t *) |
virtual void | cudaFree (void *) |
virtual void | cudaFreeHost (void *) |
virtual void * | cudaMalloc (size_t) |
virtual void * | cudaMallocHost (size_t) |
virtual void | cudaStreamWaitEvent (cudaStream_t *, cudaEvent_t *) |
virtual void | deleteCudaEvent (cudaEvent_t *) |
virtual void | deleteCudaStream (cudaStream_t *) |
virtual void | memcpyToCPU (void *, const void *, size_t, cudaStream_t *=nullptr) |
virtual void | memcpyToCUDA (void *, const void *, size_t, cudaStream_t *=nullptr) |
virtual cudaEvent_t * | newCudaEvent (bool) |
virtual cudaStream_t * | newCudaStream () |
virtual bool | streamIsActive (cudaStream_t *) |
virtual double | sumReduce (cudaStream_t *, InputArr input, size_t n)=0 |
#include <RooBatchCompute.h>
|
virtualdefault |
|
pure virtual |
Implemented in RooBatchCompute::RF_ARCH::RooBatchComputeClass.
|
pure virtual |
Implemented in RooBatchCompute::RF_ARCH::RooBatchComputeClass.
|
pure virtual |
Implemented in RooBatchCompute::RF_ARCH::RooBatchComputeClass.
|
inlinevirtual |
Definition at line 81 of file RooBatchCompute.h.
|
inlinevirtual |
Definition at line 79 of file RooBatchCompute.h.
Definition at line 71 of file RooBatchCompute.h.
Definition at line 73 of file RooBatchCompute.h.
|
inlinevirtual |
Definition at line 70 of file RooBatchCompute.h.
|
inlinevirtual |
Definition at line 72 of file RooBatchCompute.h.
|
inlinevirtual |
Definition at line 80 of file RooBatchCompute.h.
|
inlinevirtual |
Definition at line 75 of file RooBatchCompute.h.
|
inlinevirtual |
Definition at line 77 of file RooBatchCompute.h.
|
inlinevirtual |
Definition at line 83 of file RooBatchCompute.h.
|
inlinevirtual |
Definition at line 82 of file RooBatchCompute.h.
|
inlinevirtual |
Definition at line 74 of file RooBatchCompute.h.
|
inlinevirtual |
Definition at line 76 of file RooBatchCompute.h.
|
inlinevirtual |
Definition at line 78 of file RooBatchCompute.h.
|
pure virtual |
Implemented in RooBatchCompute::RF_ARCH::RooBatchComputeClass.