Caffe2 - C++ API
A deep learning, cross platform ML framework
Public Member Functions | Protected Attributes
caffe2::GPUFallbackOp< CPUOp, SkipOutputCopy > Class Template Referencefinal

A templated class to allow one to wrap a CPU operator as a CUDA operator. More...

#include <operator_fallback_gpu.h>

Inheritance diagram for caffe2::GPUFallbackOp< CPUOp, SkipOutputCopy >:
caffe2::Operator< CUDAContext > caffe2::OperatorBase caffe2::Observable< OperatorBase >

Public Member Functions

 GPUFallbackOp (const OperatorDef &def, Workspace *ws)
bool RunOnDevice () override
- Public Member Functions inherited from caffe2::Operator< CUDAContext >
 Operator (const OperatorDef &operator_def, Workspace *ws)
const Tensor< CUDAContext > & Input (int idx)
Tensor< CUDAContext > * Output (int idx)
void WaitEvent (const Event &ev, int stream_id=-1) final
void WaitEvents (const std::vector< const Event * > &events, int stream_id=-1) final
bool Run (int stream_id=0) final
bool RunAsync (int stream_id=0) final
bool IsStreamFree (int stream_id) const override
bool HasAsyncPart () const override
bool SupportsAsyncScheduling () const override
const CUDAContextgetContext () const
- Public Member Functions inherited from caffe2::OperatorBase
 OperatorBase (const OperatorDef &operator_def, Workspace *ws)
bool HasArgument (const string &name) const
 Checks if the operator has an argument of the given name.
template<typename T >
GetSingleArgument (const string &name, const T &default_value) const
template<typename T >
bool HasSingleArgumentOfType (const string &name) const
template<typename T >
vector< T > GetRepeatedArgument (const string &name, const vector< T > &default_value={}) const
template<typename T >
const T & Input (int idx)
template<typename T >
T * Output (int idx)
template<typename T >
T * Output (int idx, T *allocated)
const BlobInputBlob (int idx)
BlobOutputBlob (int idx)
template<typename T >
bool InputIsType (int idx)
template<typename T >
bool OutputIsType (int idx)
int InputSize () const
int OutputSize () const
const vector< const Blob * > & Inputs () const
const vector< Blob * > & Outputs ()
vector< TensorShape > InputTensorShapes ()
void Wait (const OperatorBase &other, int stream_id=-1)
virtual void Finish ()
virtual void AddRelatedBlobInfo (EnforceNotMet *err)
const OperatorDef & debug_def () const
void set_debug_def (const std::shared_ptr< const OperatorDef > &operator_def)
bool has_debug_def () const
void RecordLastFailedOpNetPosition ()
int net_position () const
void set_net_position (int idx)
const DeviceOption & device_option () const
const Eventevent () const
Eventevent ()
void ResetEvent ()
void DisableEvent ()
bool IsEventDisabled () const
const std::string & type () const
void annotate_engine (const std::string &engine)
const std::string & engine () const
- Public Member Functions inherited from caffe2::Observable< OperatorBase >
const ObserverAttachObserver (std::unique_ptr< Observer > observer)
std::unique_ptr< ObserverDetachObserver (const Observer *observer_ptr)
 Returns a unique_ptr to the removed observer. More...
virtual size_t NumObservers ()
void StartAllObservers ()
void StopAllObservers ()

Protected Attributes

Workspace local_ws_
vector< Blob * > local_input_blobs_
vector< Blob * > local_output_blobs_
std::unique_ptr< CPUOp > base_op_
- Protected Attributes inherited from caffe2::Operator< CUDAContext >
CUDAContext context_
- Protected Attributes inherited from caffe2::OperatorBase
std::unique_ptr< Eventevent_
- Protected Attributes inherited from caffe2::Observable< OperatorBase >
std::vector< std::unique_ptr< Observer > > observers_list_

Additional Inherited Members

- Public Types inherited from caffe2::Observable< OperatorBase >
using Observer = ObserverBase< OperatorBase >
- Static Public Attributes inherited from caffe2::OperatorBase
static constexpr int kNoNetPositionSet = -1
- Protected Member Functions inherited from caffe2::Operator< CUDAContext >
void RecordEvent (const char *err_msg=nullptr) final
std::string getErrorMsg ()
- Protected Member Functions inherited from caffe2::OperatorBase

Detailed Description

template<class CPUOp, typename SkipOutputCopy = SkipIndices<>>
class caffe2::GPUFallbackOp< CPUOp, SkipOutputCopy >

A templated class to allow one to wrap a CPU operator as a CUDA operator.

This class can be used when one does not have the CUDA implementation ready yet for an operator. Essentially, what this op does is to automatically deal with data copy for you. Plausibly, this causes a lot of overhead and is not optimal, so you should use this operator mostly for quick prototyping purpose.

All the input and output of the original operator should be TensorCPU.

Example usage: if you have a class MyMagicOp that is CPU based, and you use the registration code REGISTER_CPU_OPERATOR(MyMagic, MyMagicOp); to register the CPU side, you can create its corresponding GPU operator (with performance hits of course) via REGISTER_CUDA_OPERATOR(MyMagic, GPUFallbackOp<MyMagicOp>);

Advanced usage: if you want to have some specific outputs never copied, you can use the SkipOutputCopy template argument to do that. For example, if MyMagic produces two outputs and the first output is always going to live on the CPU, you can do REGISTER_CUDA_OPERATOR(MyMagic, GPUFallbackOp<MyMagicOp, SkipIndices<0>>);

Definition at line 40 of file operator_fallback_gpu.h.

The documentation for this class was generated from the following file: