Inheritance diagram for faiss::gpu::GpuResources:

Public Member Functions
virtual void	initializeForDevice (int device)=0

virtual cublasHandle_t	getBlasHandle (int device)=0

virtual cudaStream_t	getDefaultStream (int device)=0

virtual std::vector< cudaStream_t >	getAlternateStreams (int device)=0

virtual DeviceMemory &	getMemoryManager (int device)=0

virtual std::pair< void *, size_t >	getPinnedMemory ()=0

virtual cudaStream_t	getAsyncCopyStream (int device)=0

cublasHandle_t	getBlasHandleCurrentDevice ()

cudaStream_t	getDefaultStreamCurrentDevice ()

std::vector< cudaStream_t >	getAlternateStreamsCurrentDevice ()

DeviceMemory &	getMemoryManagerCurrentDevice ()

cudaStream_t	getAsyncCopyStreamCurrentDevice ()

Detailed Description

Base class of GPU-side resource provider; hides provision of cuBLAS handles, CUDA streams and a temporary memory manager

Definition at line 23 of file GpuResources.h.

Member Function Documentation

virtual void faiss::gpu::GpuResources::initializeForDevice ( int device )

pure virtual

Call to pre-allocate resources for a particular device. If this is not called, then resources will be allocated at the first time of demand

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