DriverAPI

Create a CUDA context. (cuCtxCreate)

Destroy a CUDA context. (cuCtxDestroy)

Disables direct access to memory allocations in a peer context and unregisters any registered allocations. (cuCtxDisablePeerAccess)

Enables direct access to memory allocations in a peer context. (cuCtxEnablePeerAccess)

Gets the context's API version. (cuCtxGetApiVersion)

Returns the CUDA context bound to the calling CPU thread. (cuCtxGetCurrent)

Returns the device ID for the current context. (cuCtxGetDevice)

Returns the flags for the current context. (cuCtxGetFlags)

Returns the unique Id associated with the context supplied. (cuCtxGetId)

Returns numerical values that correspond to the least and greatest stream priorities. (cuCtxGetStreamPriorityRange)

Pops the current CUDA context from the current CPU thread. (cuCtxPopCurrent)

Pushes a context on the current CPU thread. (cuCtxPushCurrent)

Resets all persisting lines in cache to normal status.

Binds the specified CUDA context to the calling CPU thread. (cuCtxSetCurrent)

Block for a context's tasks to complete. (cuCtxSynchronize)

Destroys a CUDA array. (cuArrayDestroy)

Destroys an external memory object. (cuDestroyExternalMemory)

Destroys an external semaphore. (cuDestroyExternalSemaphore)

Returns the default mempool of a device. (cuDeviceGetDefaultMemPool)

Gets the current mempool for a device. (cuDeviceGetMemPool)

Queries if a device may directly access a peer device's memory. (cuDeviceCanAccessPeer)

Returns a handle to a compute device. (cuDeviceGetByPCIBusId)

Returns an identifier string for the device. (cuDeviceGetName)

Returns a PCI Bus Id string for the device. (cuDeviceGetPCIBusId)

Free unused memory that was cached on the specified device for use with graphs back to the OS. (cuDeviceGraphMemTrim)

Get the state of the primary context. (cuDevicePrimaryCtxGetState)

Retain the primary context on the GPU. (cuDevicePrimaryCtxRetain)

Creates an event. (cuEventCreate)

Destroys an event. (cuEventDestroy)

Computes the elapsed time between two events. (cuEventElapsedTime)

Queries an event's status (cuEventQuery)

Records an event. (cuEventRecord)

Records an event. (cuEventRecordWithFlags)

Waits for an event to complete. (cuEventSynchronize)

Gets a handle to the embedded graph of a child graph node. (cuGraphChildGraphNodeGetGraph)

Clones a graph. (cuGraphCreate)

Creates a graph. (cuGraphCreate)

Destroys a graph. (cuGraphDestroy)

Remove a node from the graph. (cuGraphDestroyNode)

Returns the event associated with an event record node. (cuGraphEventRecordNodeGetEvent)

Sets an event record node's event. (cuGraphEventRecordNodeSetEvent)

Returns the event associated with an event wait node. (cuGraphEventWaitNodeGetEvent)

Sets an event wait node's event. (cuGraphEventWaitNodeSetEvent)

Updates node parameters in the child graph node in the given graphExec. (cuGraphExecChildGraphNodeSetParams)

Destroys an executable graph. (cuGraphExecDestroy)

Sets the event for an event record node in the given graphExec. (cuGraphExecEventRecordNodeSetEvent)

Sets the event for an event wait node in the given graphExec. (cuGraphExecEventWaitNodeSetEvent)

Returns a graph's nodes. (cuGraphGetNodes)

Set usage flags for mapping a graphics resource. (cuGraphicsResourceSetMapFlags)

Unregisters a graphics resource for access by CUDA. (cuGraphicsUnregisterResource)

Copies attributes from source node to destination node. (cuGraphKernelNodeCopyAttributes)

Launches an executable graph in a stream. (cuGraphLaunch)

Returns a memory free node's parameters. (cuGraphMemFreeNodeGetParams)

Uploads an executable graph in a stream. (cuGraphUpload)

Attempts to close memory mapped with cuIpcOpenMemHandle. (cuIpcCloseMemHandle)

Returns a function handle. (cuKernelGetFunction)

Returns a module handle. (cuLibraryGetModule)

Unloads a library. (cuLibraryUnload)

Destroys state for a JIT linker invocation. (cuLinkDestroy)

Free an address range reservation. (cuMemAddressFree)

Allocates device memory. (cuMemAlloc)

Allocates memory with stream ordered semantics. (cuMemAllocAsync)

Allocates memory that will be automatically managed by the Unified Memory system. (cuMemAllocManaged)

Frees device memory. (cuMemFree)

Frees memory with stream ordered semantics. (cuMemFreeAsync)

Frees page-locked host memory. (cuMemFreeHost)

Unregisters a memory range that was registered with cuMemHostRegister. (cuMemHostUnregister)

Destroys the specified memory pool. (cuMemPoolDestroy)

Tries to release memory back to the OS. (cuMemPoolTrimTo)

Release a memory handle representing a memory allocation which was previously allocated through cuMemCreate. (cuMemRelease)

Unmap the backing memory of a given address range. (cuMemUnmap)

Loads a compute module. (cuModuleLoad)

Unloads a module. (cuModuleUnload)

Associate a device to a multicast object. (cuMulticastAddDevice)

Attach memory to a stream asynchronously. (cuStreamAttachMemAsync)

Copies attributes from source stream to destination stream. (cuStreamCopyAttributes)

Create a stream. (cuStreamCreate)

Create a stream with the given priority. (cuStreamCreateWithPriority)

Destroys a stream. (cuStreamDestroy)

Ends capture on a stream, returning the captured graph. (cuStreamEndCapture)

Query the context associated with a stream. (cuStreamGetCtx)

Query the flags of a given stream. (cuStreamGetFlags)

Returns the unique Id associated with the stream handle supplied. (cuStreamGetId)

Query the priority of a given stream. (cuStreamGetPriority)

Determine status of a compute stream. (cuStreamQuery)

Wait until a stream's tasks are completed. (cuStreamSynchronize)

Make a compute stream wait on an event. (cuStreamWaitEvent)

Destroys a surface object. (cuSurfObjectDestroy)

Destroys a texture object. (cuTextObjectDestroy)

Release a reference to a user object. (cuUserObjectRelease)

Retain a reference to a user object. (cuUserObjectRetain)