Expand description
Provides a safe and convenient wrapper for the CUDA cuDNN API.
This crate (1.0.0) was developed against cuDNN v3.
§Architecture
This crate provides three levels of entrace.
FFI
The ffi
module exposes the foreign function interface and cuDNN specific types. Usually,
there should be no use to touch it if you only want to use cuDNN in you application. The ffi
is provided by the rust-cudnn-sys
crate and gets reexported here.
Low-Level
The api
module exposes already a complete and safe wrapper for the cuDNN API, including proper
Rust Errors. Usually there should be not need to use the API
directly though, as the Cudnn
module,
as described in the next block, provides all the API functionality but provides a more convenient interface.
High-Level
The cudnn
module exposes the Cudnn
struct, which provides a very convenient, easy-to-understand interface
for the cuDNN API. There should be not much need to obtain and read the cuDNN manual. Initialize the Cudnn
struct and you can call the available methods wich are representing all the available cuDNN operations.
§Examples
extern crate rcudnn as cudnn;
extern crate libc;
use cudnn::{Cudnn, TensorDescriptor};
use cudnn::utils::{ScalParams, DataType};
fn main() {
// Initialize a new cuDNN context and allocates resources.
let cudnn = Cudnn::new().unwrap();
// Create a cuDNN Tensor Descriptor for `src` and `dest` memory.
let src_desc = TensorDescriptor::new(&[2, 2, 2], &[4, 2, 1], DataType::Float).unwrap();
let dest_desc = TensorDescriptor::new(&[2, 2, 2], &[4, 2, 1], DataType::Float).unwrap();
let acti = cudnn.init_activation().unwrap();
// Obtain the `src` and memory pointer on the GPU.
// NOTE: You wouldn't do it like that. You need to really allocate memory on the GPU with e.g. CUDA or Collenchyma.
let src_data: *const ::libc::c_void = ::std::ptr::null();
let dest_data: *mut ::libc::c_void = ::std::ptr::null_mut();
// Now you can compute the forward sigmoid activation on your GPU.
cudnn.sigmoid_forward::<f32>(&acti, &src_desc, src_data, &dest_desc, dest_data, ScalParams::default());
}
§Notes
rust-cudnn was developed at Autumn for the Rust Machine Intelligence Framework Leaf.
rust-cudnn is part of the High-Performance Computation Framework Collenchyma, for the Neural Network Plugin. Rust CUDNN is now maintained by Juice
Modules§
Structs§
- API
- Defines the Cuda cuDNN API.
- Activation
Descriptor - Describes a ActivationDescriptor.
- CUevent_
st - CUexternal
Memory_ st - CUexternal
Semaphore_ st - CUfunc_
st - CUgraph
Exec_ st - CUgraph
Node_ st - CUgraph_
st - CUmem
Pool Handle_ st - CUstream_
st - CUuser
Object_ st - CUuuid_
st - Convolution
Descriptor - Describes a Convolution Descriptor.
- Cudnn
- Provides a the high-level interface to CUDA’s cuDNN.
- Dropout
Descriptor - Describes a DropoutDescriptor.
- Filter
Descriptor - Describes a Filter Descriptor.
- Normalization
Descriptor - Describes a LRN Descriptor.
- Pooling
Descriptor - Describes a Pooling Descriptor.
- RnnDescriptor
- Describes a Recurrent Descriptor.
- Tensor
Descriptor - Describes a TensorDescriptor.
- cuda
Access Policy Window - Specifies an access policy for a window, a contiguous extent of memory beginning at base_ptr and ending at base_ptr + num_bytes. Partition into many segments and assign segments such that. sum of “hit segments” / window == approx. ratio. sum of “miss segments” / window == approx 1-ratio. Segments and ratio specifications are fitted to the capabilities of the architecture. Accesses in a hit segment apply the hitProp access policy. Accesses in a miss segment apply the missProp access policy.
- cuda
Array - cuda
Array Sparse Properties - Sparse CUDA array and CUDA mipmapped array properties
- cuda
Array Sparse Properties__ bindgen_ ty_ 1 - cuda
Channel Format Desc - CUDA Channel format descriptor
- cuda
Device Prop - CUDA device properties
- cuda
Extent - CUDA extent
- cuda
External Memory Buffer Desc - External memory buffer descriptor
- cuda
External Memory Handle Desc - External memory handle descriptor
- cuda
External Memory Handle Desc__ bindgen_ ty_ 1__ bindgen_ ty_ 1 - Win32 handle referencing the semaphore object. Valid when type is one of the following:
- cuda
External Memory Mipmapped Array Desc - External memory mipmap descriptor
- cuda
External Semaphore Handle Desc - External semaphore handle descriptor
- cuda
External Semaphore Handle Desc__ bindgen_ ty_ 1__ bindgen_ ty_ 1 - Win32 handle referencing the semaphore object. Valid when type is one of the following:
- cuda
External Semaphore Signal Node Params - External semaphore signal node parameters
- cuda
External Semaphore Signal Params - External semaphore signal parameters, compatible with driver type
- cuda
External Semaphore Signal Params__ bindgen_ ty_ 1 - cuda
External Semaphore Signal Params__ bindgen_ ty_ 1__ bindgen_ ty_ 1 - Parameters for fence objects
- cuda
External Semaphore Signal Params__ bindgen_ ty_ 1__ bindgen_ ty_ 3 - Parameters for keyed mutex objects
- cuda
External Semaphore Wait Node Params - External semaphore wait node parameters
- cuda
External Semaphore Wait Params - External semaphore wait parameters, compatible with driver type
- cuda
External Semaphore Wait Params__ bindgen_ ty_ 1 - cuda
External Semaphore Wait Params__ bindgen_ ty_ 1__ bindgen_ ty_ 1 - Parameters for fence objects
- cuda
External Semaphore Wait Params__ bindgen_ ty_ 1__ bindgen_ ty_ 3 - Parameters for keyed mutex objects
- cuda
Func Attributes - CUDA function attributes
- cuda
Graphics Resource - CUDA graphics interop resource
- cuda
Host Node Params - CUDA host node parameters
- cuda
IpcEvent Handle_ st - CUDA IPC event handle
- cuda
IpcMem Handle_ st - CUDA IPC memory handle
- cuda
Kernel Node Params - CUDA GPU kernel node parameters
- cuda
Launch Params - CUDA launch parameters
- cuda
MemAccess Desc - Memory access descriptor
- cuda
MemAlloc Node Params - Memory allocation node parameters
- cuda
MemLocation - Specifies a memory location.
- cuda
MemPool Props - Specifies the properties of allocations made from the pool.
- cuda
MemPool PtrExport Data - Opaque data for exporting a pool allocation
- cuda
Memcpy3D Parms - CUDA 3D memory copying parameters
- cuda
Memcpy3D Peer Parms - CUDA 3D cross-device memory copying parameters
- cuda
Memset Params - CUDA Memset node parameters
- cuda
Mipmapped Array - cuda
Pitched Ptr - CUDA Pitched memory pointer
- cuda
Pointer Attributes - CUDA pointer attributes
- cudaPos
- CUDA 3D position
- cuda
Resource Desc - CUDA resource descriptor
- cuda
Resource Desc__ bindgen_ ty_ 1__ bindgen_ ty_ 1 - cuda
Resource Desc__ bindgen_ ty_ 1__ bindgen_ ty_ 2 - cuda
Resource Desc__ bindgen_ ty_ 1__ bindgen_ ty_ 3 - cuda
Resource Desc__ bindgen_ ty_ 1__ bindgen_ ty_ 4 - cuda
Resource View Desc - CUDA resource view descriptor
- cuda
Texture Desc - CUDA texture descriptor
- cudnn
Activation Struct - cudnn
Algorithm Performance Struct - cudnn
Algorithm Struct - cudnn
Algorithm Union Struct - cudnn
Attn Struct - cudnnCTC
Loss Struct - cudnn
Context - cudnn
Convolution BwdData Algo Perf Struct - cudnn
Convolution BwdFilter Algo Perf Struct - cudnn
Convolution FwdAlgo Perf Struct - cudnn
Convolution Struct - cudnn
Debug Struct - cudnn
Dropout Struct - cudnn
Filter Struct - cudnn
Fused OpsConst Param Struct - cudnn
Fused OpsPlan Struct - cudnn
Fused OpsVariant Param Struct - cudnnLRN
Struct - cudnn
OpTensor Struct - cudnn
PersistentRNN Plan - cudnn
Pooling Struct - cudnnRNN
Data Struct - cudnnRNN
Struct - cudnn
Reduce Tensor Struct - cudnn
Runtime Tag_ t - cudnn
SeqData Struct - cudnn
Spatial Transformer Struct - cudnn
Tensor Struct - cudnn
Tensor Transform Struct - dim3
- surface
Reference - CUDA Surface reference
- texture
Reference - CUDA texture reference
Enums§
- Error
- Defines CUDA’s cuDNN errors.
- cuda
Access Property - Specifies performance hint with ::cudaAccessPolicyWindow for hitProp and missProp members.
- cudaCG
Scope - CUDA cooperative group scope
- cuda
Channel Format Kind - Channel format kind
- cuda
Compute Mode - CUDA device compute modes
- cuda
Data Type - cuda
Data Type_ t - cuda
Device Attr - CUDA device attributes
- cuda
Device P2PAttr - CUDA device P2P attributes
- cuda
Error - CUDA error types
- cuda
Error_ t - CUDA error types
- cuda
External Memory Handle Type - External memory handle types
- cuda
External Semaphore Handle Type - External semaphore handle types
- cuda
FlushGPU DirectRDMA Writes Options - CUDA GPUDirect RDMA flush writes APIs supported on the device
- cuda
FlushGPU DirectRDMA Writes Scope - CUDA GPUDirect RDMA flush writes scopes
- cuda
FlushGPU DirectRDMA Writes Target - CUDA GPUDirect RDMA flush writes targets
- cuda
Func Attribute - CUDA function attributes that can be set using ::cudaFuncSetAttribute
- cuda
Func Cache - CUDA function cache configurations
- cudaGPU
DirectRDMA Writes Ordering - CUDA GPUDirect RDMA flush writes ordering features of the device
- cuda
GetDriver Entry Point Flags - Flags to specify search options to be used with ::cudaGetDriverEntryPoint For more details see ::cuGetProcAddress
- cuda
Graph Debug DotFlags - CUDA Graph debug write options
- cuda
Graph Exec Update Result - CUDA Graph Update error types
- cuda
Graph Instantiate Flags - Flags for instantiating a graph
- cuda
Graph MemAttribute Type - Graph memory attributes
- cuda
Graph Node Type - CUDA Graph node types
- cuda
Graphics Cube Face - CUDA graphics interop array indices for cube maps
- cuda
Graphics MapFlags - CUDA graphics interop map flags
- cuda
Graphics Register Flags - CUDA graphics interop register flags
- cuda
Kernel Node AttrID - Graph kernel node Attributes
- cuda
Limit - CUDA Limits
- cuda
MemAccess Flags - Specifies the memory protection flags for mapping.
- cuda
MemAllocation Handle Type - Flags for specifying particular handle types
- cuda
MemAllocation Type - Defines the allocation types available
- cuda
MemLocation Type - Specifies the type of location
- cuda
MemPool Attr - CUDA memory pool attributes
- cuda
MemRange Attribute - CUDA range attributes
- cuda
Memcpy Kind - CUDA memory copy types
- cuda
Memory Advise - CUDA Memory Advise values
- cuda
Memory Type - CUDA memory types
- cuda
Output Mode - CUDA Profiler Output modes
- cuda
Output Mode_ t - CUDA Profiler Output modes
- cuda
Resource Type - CUDA resource types
- cuda
Resource View Format - CUDA texture resource view formats
- cuda
Round Mode - cuda
Shared Carveout - Shared memory carveout configurations. These may be passed to cudaFuncSetAttribute
- cuda
Shared MemConfig - CUDA shared memory configuration
- cuda
Stream AttrID - Stream Attributes
- cuda
Stream Capture Mode - Possible modes for stream capture thread interactions. For more details see ::cudaStreamBeginCapture and ::cudaThreadExchangeStreamCaptureMode
- cuda
Stream Capture Status - Possible stream capture statuses returned by ::cudaStreamIsCapturing
- cuda
Stream Update Capture Dependencies Flags - Flags for ::cudaStreamUpdateCaptureDependencies
- cuda
Surface Boundary Mode - CUDA Surface boundary modes
- cuda
Surface Format Mode - CUDA Surface format modes
- cuda
Synchronization Policy - cuda
Texture Address Mode - CUDA texture address modes
- cuda
Texture Filter Mode - CUDA texture filter modes
- cuda
Texture Read Mode - CUDA texture read modes
- cuda
User Object Flags - Flags for user objects for graphs
- cuda
User Object Retain Flags - Flags for retaining user object references for graphs
- cudnn
Activation Mode_ t - cudnn
Backend Attribute Name_ t - cudnn
Backend Attribute Type_ t - cudnn
Backend Behavior Note_ t - cudnn
Backend Descriptor Type_ t - cudnn
Backend Heur Mode_ t - cudnn
Backend Knob Type_ t - cudnn
Backend Layout Type_ t - cudnn
Backend Numerical Note_ t - cudnn
Batch Norm Mode_ t - cudnn
Batch Norm Ops_ t - cudnn
BnFinalize Stats Mode_ t - cudnnCTC
Loss Algo_ t - cudnn
Convolution BwdData Algo_ t - cudnn
Convolution BwdFilter Algo_ t - cudnn
Convolution FwdAlgo_ t - cudnn
Convolution Mode_ t - cudnn
Data Type_ t - cudnn
Determinism_ t - cudnn
Direction Mode_ t - cudnn
DivNorm Mode_ t - cudnn
ErrQuery Mode_ t - cudnn
Folding Direction_ t - cudnn
Forward Mode_ t - cudnn
Fused OpsConst Param Label_ t - cudnn
Fused OpsPointer Place Holder_ t - cudnn
Fused OpsVariant Param Label_ t - cudnn
Fused Ops_ t - cudnn
GenStats Mode_ t - cudnn
Indices Type_ t - cudnnLRN
Mode_ t - cudnn
Loss Normalization Mode_ t - cudnn
Math Type_ t - cudnn
Multi Head Attn Weight Kind_ t - cudnn
NanPropagation_ t - cudnn
Norm Algo_ t - cudnn
Norm Mode_ t - cudnn
Norm Ops_ t - cudnn
OpTensor Op_ t - cudnn
Pointwise Mode_ t - cudnn
Pooling Mode_ t - cudnnRNN
Algo_ t - cudnnRNN
Bias Mode_ t - cudnnRNN
Clip Mode_ t - cudnnRNN
Data Layout_ t - cudnnRNN
Input Mode_ t - cudnnRNN
Mode_ t - cudnn
Reduce Tensor Indices_ t - cudnn
Reduce Tensor Op_ t - cudnn
Reorder Type_ t - cudnn
Sampler Type_ t - cudnn
SeqData Axis_ t - cudnn
Severity_ t - cudnn
Softmax Algorithm_ t - cudnn
Softmax Mode_ t - cudnn
Status_ t - cudnn
Tensor Format_ t - cudnn
Wgrad Mode_ t
Constants§
- CUDNN_
ADV_ INFER_ MAJOR - CUDNN_
ADV_ INFER_ MINOR - CUDNN_
ADV_ INFER_ PATCH - CUDNN_
ADV_ TRAIN_ MAJOR - CUDNN_
ADV_ TRAIN_ MINOR - CUDNN_
ADV_ TRAIN_ PATCH - CUDNN_
ATTN_ DISABLE_ PROJ_ BIASES - CUDNN_
ATTN_ ENABLE_ PROJ_ BIASES - CUDNN_
ATTN_ QUERYMAP_ ALL_ TO_ ONE - CUDNN_
ATTN_ QUERYMAP_ ONE_ TO_ ONE - CUDNN_
ATTN_ WKIND_ COUNT - CUDNN_
BN_ MIN_ EPSILON - CUDNN_
CNN_ INFER_ MAJOR - CUDNN_
CNN_ INFER_ MINOR - CUDNN_
CNN_ INFER_ PATCH - CUDNN_
CNN_ TRAIN_ MAJOR - CUDNN_
CNN_ TRAIN_ MINOR - CUDNN_
CNN_ TRAIN_ PATCH - CUDNN_
DIM_ MAX - CUDNN_
LRN_ MAX_ N - CUDNN_
LRN_ MIN_ BETA - CUDNN_
LRN_ MIN_ K - CUDNN_
LRN_ MIN_ N - CUDNN_
MAJOR - CUDNN_
MINOR - CUDNN_
OPS_ INFER_ MAJOR - CUDNN_
OPS_ INFER_ MINOR - CUDNN_
OPS_ INFER_ PATCH - CUDNN_
OPS_ TRAIN_ MAJOR - CUDNN_
OPS_ TRAIN_ MINOR - CUDNN_
OPS_ TRAIN_ PATCH - CUDNN_
PATCHLEVEL - CUDNN_
RNN_ PADDED_ IO_ DISABLED - CUDNN_
RNN_ PADDED_ IO_ ENABLED - CUDNN_
SEQDATA_ DIM_ COUNT - CUDNN_
VERSION - library
Property Type_ t_ MAJOR_ VERSION - library
Property Type_ t_ MINOR_ VERSION - library
Property Type_ t_ PATCH_ LEVEL
Functions§
- cuda
Array ⚠GetInfo - \brief Gets info about the specified cudaArray
- cuda
Array ⚠GetPlane - \brief Gets a CUDA array plane from a CUDA array
- cuda
Array ⚠GetSparse Properties - cuda
Bind ⚠Surface ToArray - \brief Binds an array to a surface
- cuda
Bind ⚠Texture - \brief Binds a memory area to a texture
- cuda
Bind ⚠Texture2D - \brief Binds a 2D memory area to a texture
- cuda
Bind ⚠Texture ToArray - \brief Binds an array to a texture
- cuda
Bind ⚠Texture ToMipmapped Array - \brief Binds a mipmapped array to a texture
- cuda
Choose ⚠Device - \brief Select compute-device which best matches criteria
- cuda
Create ⚠Channel Desc - \brief Returns a channel descriptor using the specified format
- cuda
Create ⚠Surface Object - \brief Creates a surface object
- cuda
Create ⚠Texture Object - \brief Creates a texture object
- cuda
CtxReset ⚠Persisting L2Cache - \brief Resets all persisting lines in cache to normal status.
- cuda
Destroy ⚠External Memory - \brief Destroys an external memory object.
- cuda
Destroy ⚠External Semaphore - \brief Destroys an external semaphore
- cuda
Destroy ⚠Surface Object - \brief Destroys a surface object
- cuda
Destroy ⚠Texture Object - \brief Destroys a texture object
- cuda
Device ⚠CanAccess Peer - \brief Queries if a device may directly access a peer device’s memory.
- cuda
Device ⚠Disable Peer Access - \brief Disables direct access to memory allocations on a peer device.
- cuda
Device ⚠Enable Peer Access - \brief Enables direct access to memory allocations on a peer device.
- cuda
Device ⚠FlushGPU DirectRDMA Writes - cuda
Device ⚠GetAttribute - \brief Returns information about the device
- cuda
Device ⚠GetByPCI BusId - \brief Returns a handle to a compute device
- cuda
Device ⚠GetCache Config - \brief Returns the preferred cache configuration for the current device.
- cuda
Device ⚠GetDefault MemPool - \brief Returns the default mempool of a device
- cuda
Device ⚠GetGraph MemAttribute - cuda
Device ⚠GetLimit - \brief Returns resource limits
- cuda
Device ⚠GetMem Pool - \brief Gets the current mempool for a device
- cuda
Device ⚠GetNv SciSync Attributes - \brief Return NvSciSync attributes that this device can support.
- cuda
Device ⚠GetP2P Attribute - \brief Queries attributes of the link between two devices.
- cuda
Device ⚠GetPCI BusId - \brief Returns a PCI Bus Id string for the device
- cuda
Device ⚠GetShared MemConfig - \brief Returns the shared memory configuration for the current device.
- cuda
Device ⚠GetStream Priority Range - \brief Returns numerical values that correspond to the least and greatest stream priorities.
- cuda
Device ⚠GetTexture1D Linear MaxWidth - cuda
Device ⚠Graph MemTrim - cuda
Device ⚠Reset - \brief Destroy all allocations and reset all state on the current device in the current process.
- cuda
Device ⚠SetCache Config - \brief Sets the preferred cache configuration for the current device.
- cuda
Device ⚠SetGraph MemAttribute - cuda
Device ⚠SetLimit - \brief Set resource limits
- cuda
Device ⚠SetMem Pool - \brief Sets the current memory pool of a device
- cuda
Device ⚠SetShared MemConfig - \brief Sets the shared memory configuration for the current device.
- cuda
Device ⚠Synchronize - \brief Wait for compute device to finish
- cuda
Driver ⚠GetVersion - \brief Returns the latest version of CUDA supported by the driver
- cuda
Event ⚠Create - \brief Creates an event object
- cuda
Event ⚠Create With Flags - \brief Creates an event object with the specified flags
- cuda
Event ⚠Destroy - \brief Destroys an event object
- cuda
Event ⚠Elapsed Time - \brief Computes the elapsed time between events
- cuda
Event ⚠Query - \brief Queries an event’s status
- cuda
Event ⚠Record - \brief Records an event
- cuda
Event ⚠Record With Flags - cuda
Event ⚠Synchronize - \brief Waits for an event to complete
- cuda
External ⚠Memory GetMapped Buffer - \brief Maps a buffer onto an imported memory object
- cuda
External ⚠Memory GetMapped Mipmapped Array - \brief Maps a CUDA mipmapped array onto an external memory object
- cuda
Free ⚠ - \brief Frees memory on the device
- cuda
Free ⚠Array - \brief Frees an array on the device
- cuda
Free ⚠Async - \brief Frees memory with stream ordered semantics
- cuda
Free ⚠Host - \brief Frees page-locked memory
- cuda
Free ⚠Mipmapped Array - \brief Frees a mipmapped array on the device
- cuda
Func ⚠GetAttributes - \brief Find out attributes for a given function
- cuda
Func ⚠SetAttribute - \brief Set attributes for a given function
- cuda
Func ⚠SetCache Config - \brief Sets the preferred cache configuration for a device function
- cuda
Func ⚠SetShared MemConfig - \brief Sets the shared memory configuration for a device function
- cuda
GetChannel ⚠Desc - \brief Get the channel descriptor of an array
- cuda
GetDevice ⚠ - \brief Returns which device is currently being used
- cuda
GetDevice ⚠Count - \brief Returns the number of compute-capable devices
- cuda
GetDevice ⚠Flags - \brief Gets the flags for the current device
- cuda
GetDevice ⚠Properties - \brief Returns information about the compute-device
- cuda
GetDriver ⚠Entry Point - \brief Returns the requested driver API function pointer
- cuda
GetError ⚠Name - \brief Returns the string representation of an error code enum name
- cuda
GetError ⚠String - \brief Returns the description string for an error code
- cuda
GetExport ⚠Table - \cond impl_private
- cuda
GetFunc ⚠BySymbol - \brief Get pointer to device entry function that matches entry function \p symbolPtr
- cuda
GetLast ⚠Error - \brief Returns the last error from a runtime call
- cuda
GetMipmapped ⚠Array Level - \brief Gets a mipmap level of a CUDA mipmapped array
- cuda
GetSurface ⚠Object Resource Desc - \brief Returns a surface object’s resource descriptor Returns the resource descriptor for the surface object specified by \p surfObject.
- cuda
GetSurface ⚠Reference - \brief Get the surface reference associated with a symbol
- cuda
GetSymbol ⚠Address - \brief Finds the address associated with a CUDA symbol
- cuda
GetSymbol ⚠Size - \brief Finds the size of the object associated with a CUDA symbol
- cuda
GetTexture ⚠Alignment Offset - \brief Get the alignment offset of a texture
- cuda
GetTexture ⚠Object Resource Desc - \brief Returns a texture object’s resource descriptor
- cuda
GetTexture ⚠Object Resource View Desc - \brief Returns a texture object’s resource view descriptor
- cuda
GetTexture ⚠Object Texture Desc - \brief Returns a texture object’s texture descriptor
- cuda
GetTexture ⚠Reference - \brief Get the texture reference associated with a symbol
- cuda
Graph ⚠AddChild Graph Node - \brief Creates a child graph node and adds it to a graph
- cuda
Graph ⚠AddDependencies - \brief Adds dependency edges to a graph.
- cuda
Graph ⚠AddEmpty Node - \brief Creates an empty node and adds it to a graph
- cuda
Graph ⚠AddEvent Record Node - cuda
Graph ⚠AddEvent Wait Node - cuda
Graph ⚠AddExternal Semaphores Signal Node - cuda
Graph ⚠AddExternal Semaphores Wait Node - cuda
Graph ⚠AddHost Node - \brief Creates a host execution node and adds it to a graph
- cuda
Graph ⚠AddKernel Node - \brief Creates a kernel execution node and adds it to a graph
- cuda
Graph ⚠AddMem Alloc Node - cuda
Graph ⚠AddMem Free Node - cuda
Graph ⚠AddMemcpy Node - \brief Creates a memcpy node and adds it to a graph
- cuda
Graph ⚠AddMemcpy Node1D - cuda
Graph ⚠AddMemcpy Node From Symbol - cuda
Graph ⚠AddMemcpy Node ToSymbol - cuda
Graph ⚠AddMemset Node - \brief Creates a memset node and adds it to a graph
- cuda
Graph ⚠Child Graph Node GetGraph - \brief Gets a handle to the embedded graph of a child graph node
- cuda
Graph ⚠Clone - \brief Clones a graph
- cuda
Graph ⚠Create - \brief Creates a graph
- cuda
Graph ⚠Debug DotPrint - \brief Write a DOT file describing graph structure
- cuda
Graph ⚠Destroy - \brief Destroys a graph
- cuda
Graph ⚠Destroy Node - \brief Remove a node from the graph
- cuda
Graph ⚠Event Record Node GetEvent - cuda
Graph ⚠Event Record Node SetEvent - cuda
Graph ⚠Event Wait Node GetEvent - cuda
Graph ⚠Event Wait Node SetEvent - cuda
Graph ⚠Exec Child Graph Node SetParams - cuda
Graph ⚠Exec Destroy - \brief Destroys an executable graph
- cuda
Graph ⚠Exec Event Record Node SetEvent - cuda
Graph ⚠Exec Event Wait Node SetEvent - cuda
Graph ⚠Exec External Semaphores Signal Node SetParams - cuda
Graph ⚠Exec External Semaphores Wait Node SetParams - cuda
Graph ⚠Exec Host Node SetParams - \brief Sets the parameters for a host node in the given graphExec.
- cuda
Graph ⚠Exec Kernel Node SetParams - \brief Sets the parameters for a kernel node in the given graphExec
- cuda
Graph ⚠Exec Memcpy Node SetParams - \brief Sets the parameters for a memcpy node in the given graphExec.
- cuda
Graph ⚠Exec Memcpy Node SetParams1D - cuda
Graph ⚠Exec Memcpy Node SetParams From Symbol - cuda
Graph ⚠Exec Memcpy Node SetParams ToSymbol - cuda
Graph ⚠Exec Memset Node SetParams - \brief Sets the parameters for a memset node in the given graphExec.
- cuda
Graph ⚠Exec Update - \brief Check whether an executable graph can be updated with a graph and perform the update if possible
- cuda
Graph ⚠External Semaphores Signal Node GetParams - cuda
Graph ⚠External Semaphores Signal Node SetParams - cuda
Graph ⚠External Semaphores Wait Node GetParams - cuda
Graph ⚠External Semaphores Wait Node SetParams - cuda
Graph ⚠GetEdges - \brief Returns a graph’s dependency edges
- cuda
Graph ⚠GetNodes - \brief Returns a graph’s nodes
- cuda
Graph ⚠GetRoot Nodes - \brief Returns a graph’s root nodes
- cuda
Graph ⚠Host Node GetParams - \brief Returns a host node’s parameters
- cuda
Graph ⚠Host Node SetParams - \brief Sets a host node’s parameters
- cuda
Graph ⚠Instantiate - \brief Creates an executable graph from a graph
- cuda
Graph ⚠Instantiate With Flags - cuda
Graph ⚠Kernel Node Copy Attributes - \brief Copies attributes from source node to destination node.
- cuda
Graph ⚠Kernel Node GetAttribute - \brief Queries node attribute.
- cuda
Graph ⚠Kernel Node GetParams - \brief Returns a kernel node’s parameters
- cuda
Graph ⚠Kernel Node SetAttribute - \brief Sets node attribute.
- cuda
Graph ⚠Kernel Node SetParams - \brief Sets a kernel node’s parameters
- cuda
Graph ⚠Launch - \brief Launches an executable graph in a stream
- cuda
Graph ⚠MemAlloc Node GetParams - cuda
Graph ⚠MemFree Node GetParams - cuda
Graph ⚠Memcpy Node GetParams - \brief Returns a memcpy node’s parameters
- cuda
Graph ⚠Memcpy Node SetParams - \brief Sets a memcpy node’s parameters
- cuda
Graph ⚠Memcpy Node SetParams1D - cuda
Graph ⚠Memcpy Node SetParams From Symbol - cuda
Graph ⚠Memcpy Node SetParams ToSymbol - cuda
Graph ⚠Memset Node GetParams - \brief Returns a memset node’s parameters
- cuda
Graph ⚠Memset Node SetParams - \brief Sets a memset node’s parameters
- cuda
Graph ⚠Node Find InClone - \brief Finds a cloned version of a node
- cuda
Graph ⚠Node GetDependencies - \brief Returns a node’s dependencies
- cuda
Graph ⚠Node GetDependent Nodes - \brief Returns a node’s dependent nodes
- cuda
Graph ⚠Node GetType - \brief Returns a node’s type
- cuda
Graph ⚠Release User Object - \brief Release a user object reference from a graph
- cuda
Graph ⚠Remove Dependencies - \brief Removes dependency edges from a graph.
- cuda
Graph ⚠Retain User Object - \brief Retain a reference to a user object from a graph
- cuda
Graph ⚠Upload - cuda
Graphics ⚠MapResources - \brief Map graphics resources for access by CUDA
- cuda
Graphics ⚠Resource GetMapped Mipmapped Array - \brief Get a mipmapped array through which to access a mapped graphics resource.
- cuda
Graphics ⚠Resource GetMapped Pointer - \brief Get an device pointer through which to access a mapped graphics resource.
- cuda
Graphics ⚠Resource SetMap Flags - \brief Set usage flags for mapping a graphics resource
- cuda
Graphics ⚠SubResource GetMapped Array - \brief Get an array through which to access a subresource of a mapped graphics resource.
- cuda
Graphics ⚠Unmap Resources - \brief Unmap graphics resources.
- cuda
Graphics ⚠Unregister Resource - \brief Unregisters a graphics resource for access by CUDA
- cuda
Host ⚠Alloc - \brief Allocates page-locked memory on the host
- cuda
Host ⚠GetDevice Pointer - \brief Passes back device pointer of mapped host memory allocated by cudaHostAlloc or registered by cudaHostRegister
- cuda
Host ⚠GetFlags - \brief Passes back flags used to allocate pinned host memory allocated by cudaHostAlloc
- cuda
Host ⚠Register - \brief Registers an existing host memory range for use by CUDA
- cuda
Host ⚠Unregister - \brief Unregisters a memory range that was registered with cudaHostRegister
- cuda
Import ⚠External Memory - \brief Imports an external memory object
- cuda
Import ⚠External Semaphore - \brief Imports an external semaphore
- cuda
IpcClose ⚠MemHandle - \brief Attempts to close memory mapped with cudaIpcOpenMemHandle
- cuda
IpcGet ⚠Event Handle - \brief Gets an interprocess handle for a previously allocated event
- cuda
IpcGet ⚠MemHandle - \brief Gets an interprocess memory handle for an existing device memory allocation
- cuda
IpcOpen ⚠Event Handle - \brief Opens an interprocess event handle for use in the current process
- cuda
IpcOpen ⚠MemHandle - \brief Opens an interprocess memory handle exported from another process and returns a device pointer usable in the local process.
- cuda
Launch ⚠Cooperative Kernel - \brief Launches a device function where thread blocks can cooperate and synchronize as they execute
- cuda
Launch ⚠Cooperative Kernel Multi Device - \brief Launches device functions on multiple devices where thread blocks can cooperate and synchronize as they execute
- cuda
Launch ⚠Host Func - \brief Enqueues a host function call in a stream
- cuda
Launch ⚠Kernel - \brief Launches a device function
- cuda
Malloc ⚠ - \brief Allocate memory on the device
- cuda
Malloc3D ⚠ - \brief Allocates logical 1D, 2D, or 3D memory objects on the device
- cuda
Malloc3D ⚠Array - \brief Allocate an array on the device
- cuda
Malloc ⚠Array - \brief Allocate an array on the device
- cuda
Malloc ⚠Async - \brief Allocates memory with stream ordered semantics
- cuda
Malloc ⚠From Pool Async - \brief Allocates memory from a specified pool with stream ordered semantics.
- cuda
Malloc ⚠Host - \brief Allocates page-locked memory on the host
- cuda
Malloc ⚠Managed - cuda
Malloc ⚠Mipmapped Array - \brief Allocate a mipmapped array on the device
- cuda
Malloc ⚠Pitch - \brief Allocates pitched memory on the device
- cuda
MemAdvise ⚠ - \brief Advise about the usage of a given memory range
- cuda
MemGet ⚠Info - \brief Gets free and total device memory
- cuda
MemPool ⚠Create - \brief Creates a memory pool
- cuda
MemPool ⚠Destroy - \brief Destroys the specified memory pool
- cuda
MemPool ⚠Export Pointer - \brief Export data to share a memory pool allocation between processes.
- cuda
MemPool ⚠Export ToShareable Handle - \brief Exports a memory pool to the requested handle type.
- cuda
MemPool ⚠GetAccess - \brief Returns the accessibility of a pool from a device
- cuda
MemPool ⚠GetAttribute - \brief Gets attributes of a memory pool
- cuda
MemPool ⚠Import From Shareable Handle - \brief imports a memory pool from a shared handle.
- cuda
MemPool ⚠Import Pointer - \brief Import a memory pool allocation from another process.
- cuda
MemPool ⚠SetAccess - \brief Controls visibility of pools between devices
- cuda
MemPool ⚠SetAttribute - \brief Sets attributes of a memory pool
- cuda
MemPool ⚠Trim To - \brief Tries to release memory back to the OS
- cuda
MemPrefetch ⚠Async - \brief Prefetches memory to the specified destination device
- cuda
MemRange ⚠GetAttribute - \brief Query an attribute of a given memory range
- cuda
MemRange ⚠GetAttributes - \brief Query attributes of a given memory range.
- cuda
Memcpy ⚠ - \brief Copies data between host and device
- cuda
Memcpy2D ⚠ - \brief Copies data between host and device
- cuda
Memcpy2D ⚠Array ToArray - \brief Copies data between host and device
- cuda
Memcpy2D ⚠Async - \brief Copies data between host and device
- cuda
Memcpy2D ⚠From Array - \brief Copies data between host and device
- cuda
Memcpy2D ⚠From Array Async - \brief Copies data between host and device
- cuda
Memcpy2D ⚠ToArray - \brief Copies data between host and device
- cuda
Memcpy2D ⚠ToArray Async - \brief Copies data between host and device
- cuda
Memcpy3D ⚠ - \brief Copies data between 3D objects
- cuda
Memcpy3D ⚠Async - \brief Copies data between 3D objects
- cuda
Memcpy3D ⚠Peer - \brief Copies memory between devices
- cuda
Memcpy3D ⚠Peer Async - \brief Copies memory between devices asynchronously.
- cuda
Memcpy ⚠Array ToArray - \brief Copies data between host and device
- cuda
Memcpy ⚠Async - \brief Copies data between host and device
- cuda
Memcpy ⚠From Array - \brief Copies data between host and device
- cuda
Memcpy ⚠From Array Async - \brief Copies data between host and device
- cuda
Memcpy ⚠From Symbol - \brief Copies data from the given symbol on the device
- cuda
Memcpy ⚠From Symbol Async - \brief Copies data from the given symbol on the device
- cuda
Memcpy ⚠Peer - \brief Copies memory between two devices
- cuda
Memcpy ⚠Peer Async - \brief Copies memory between two devices asynchronously.
- cuda
Memcpy ⚠ToArray - \brief Copies data between host and device
- cuda
Memcpy ⚠ToArray Async - \brief Copies data between host and device
- cuda
Memcpy ⚠ToSymbol - \brief Copies data to the given symbol on the device
- cuda
Memcpy ⚠ToSymbol Async - \brief Copies data to the given symbol on the device
- cuda
Memset ⚠ - \brief Initializes or sets device memory to a value
- cuda
Memset2D ⚠ - \brief Initializes or sets device memory to a value
- cuda
Memset2D ⚠Async - \brief Initializes or sets device memory to a value
- cuda
Memset3D ⚠ - \brief Initializes or sets device memory to a value
- cuda
Memset3D ⚠Async - \brief Initializes or sets device memory to a value
- cuda
Memset ⚠Async - \brief Initializes or sets device memory to a value
- cuda
Mipmapped ⚠Array GetSparse Properties - cuda
Occupancy ⚠Available DynamicS MemPer Block - \brief Returns dynamic shared memory available per block when launching \p numBlocks blocks on SM.
- cuda
Occupancy ⚠MaxActive Blocks PerMultiprocessor - \brief Returns occupancy for a device function
- cuda
Occupancy ⚠MaxActive Blocks PerMultiprocessor With Flags - \brief Returns occupancy for a device function with the specified flags
- cuda
Peek ⚠AtLast Error - \brief Returns the last error from a runtime call
- cuda
Pointer ⚠GetAttributes - \brief Returns attributes about a specified pointer
- cuda
Runtime ⚠GetVersion - \brief Returns the CUDA Runtime version
- cuda
SetDevice ⚠ - \brief Set device to be used for GPU executions
- cuda
SetDevice ⚠Flags - \brief Sets flags to be used for device executions
- cuda
SetDouble ⚠ForDevice - \brief Converts a double argument to be executed on a device
- cuda
SetDouble ⚠ForHost - \brief Converts a double argument after execution on a device
- cuda
SetValid ⚠Devices - \brief Set a list of devices that can be used for CUDA
- cuda
Signal ⚠External Semaphores Async_ v2 - cuda
Stream ⚠AddCallback - \brief Add a callback to a compute stream
- cuda
Stream ⚠Attach MemAsync - cuda
Stream ⚠Begin Capture - \brief Begins graph capture on a stream
- cuda
Stream ⚠Copy Attributes - \brief Copies attributes from source stream to destination stream.
- cuda
Stream ⚠Create - \brief Create an asynchronous stream
- cuda
Stream ⚠Create With Flags - \brief Create an asynchronous stream
- cuda
Stream ⚠Create With Priority - \brief Create an asynchronous stream with the specified priority
- cuda
Stream ⚠Destroy - \brief Destroys and cleans up an asynchronous stream
- cuda
Stream ⚠EndCapture - \brief Ends capture on a stream, returning the captured graph
- cuda
Stream ⚠GetAttribute - \brief Queries stream attribute.
- cuda
Stream ⚠GetCapture Info - \brief Query capture status of a stream
- cuda
Stream ⚠GetCapture Info_ v2 - \brief Query a stream’s capture state (11.3+)
- cuda
Stream ⚠GetFlags - \brief Query the flags of a stream
- cuda
Stream ⚠GetPriority - \brief Query the priority of a stream
- cuda
Stream ⚠IsCapturing - \brief Returns a stream’s capture status
- cuda
Stream ⚠Query - \brief Queries an asynchronous stream for completion status
- cuda
Stream ⚠SetAttribute - \brief Sets stream attribute.
- cuda
Stream ⚠Synchronize - \brief Waits for stream tasks to complete
- cuda
Stream ⚠Update Capture Dependencies - \brief Update the set of dependencies in a capturing stream (11.3+)
- cuda
Stream ⚠Wait Event - \brief Make a compute stream wait on an event
- cuda
Thread ⚠Exchange Stream Capture Mode - \brief Swaps the stream capture interaction mode for a thread
- cuda
Thread ⚠Exit - \brief Exit and clean up from CUDA launches
- cuda
Thread ⚠GetCache Config - \brief Returns the preferred cache configuration for the current device.
- cuda
Thread ⚠GetLimit - \brief Returns resource limits
- cuda
Thread ⚠SetCache Config - \brief Sets the preferred cache configuration for the current device.
- cuda
Thread ⚠SetLimit - \brief Set resource limits
- cuda
Thread ⚠Synchronize - \brief Wait for compute device to finish
- cuda
Unbind ⚠Texture - \brief Unbinds a texture
- cuda
User ⚠Object Create - \brief Create a user object
- cuda
User ⚠Object Release - \brief Release a reference to a user object
- cuda
User ⚠Object Retain - \brief Retain a reference to a user object
- cuda
Wait ⚠External Semaphores Async_ v2 - cudnn
Activation ⚠Backward - cudnn
Activation ⚠Forward - cudnn
AddTensor ⚠ - cudnn
AdvInfer ⚠Version Check - cudnn
AdvTrain ⚠Version Check - cudnn
Backend ⚠Create Descriptor - cudnn
Backend ⚠Destroy Descriptor - cudnn
Backend ⚠Execute - cudnn
Backend ⚠Finalize - cudnn
Backend ⚠GetAttribute - cudnn
Backend ⚠Initialize - cudnn
Backend ⚠SetAttribute - cudnn
Batch ⚠Normalization Backward - cudnn
Batch ⚠Normalization Backward Ex - cudnn
Batch ⚠Normalization Forward Inference - cudnn
Batch ⚠Normalization Forward Training - cudnn
Batch ⚠Normalization Forward Training Ex - cudnn
BuildRNN ⚠Dynamic - cudnnCTC
Loss ⚠ - cudnnCTC
Loss_ ⚠v8 - cudnn
CnnInfer ⚠Version Check - cudnn
CnnTrain ⚠Version Check - cudnn
Convolution ⚠Backward Bias - cudnn
Convolution ⚠Backward Data - cudnn
Convolution ⚠Backward Filter - cudnn
Convolution ⚠Bias Activation Forward - cudnn
Convolution ⚠Forward - cudnn
Copy ⚠Algorithm Descriptor - cudnn
Create ⚠ - cudnn
Create ⚠Activation Descriptor - cudnn
Create ⚠Algorithm Descriptor - cudnn
Create ⚠Algorithm Performance - cudnn
Create ⚠Attn Descriptor - cudnn
CreateCTC ⚠Loss Descriptor - cudnn
Create ⚠Convolution Descriptor - cudnn
Create ⚠Dropout Descriptor - cudnn
Create ⚠Filter Descriptor - cudnn
Create ⚠Fused OpsConst Param Pack - cudnn
Create ⚠Fused OpsPlan - cudnn
Create ⚠Fused OpsVariant Param Pack - cudnn
CreateLRN ⚠Descriptor - cudnn
Create ⚠OpTensor Descriptor - cudnn
Create ⚠PersistentRNN Plan - cudnn
Create ⚠Pooling Descriptor - cudnn
CreateRNN ⚠Data Descriptor - cudnn
CreateRNN ⚠Descriptor - cudnn
Create ⚠Reduce Tensor Descriptor - cudnn
Create ⚠SeqData Descriptor - cudnn
Create ⚠Spatial Transformer Descriptor - cudnn
Create ⚠Tensor Descriptor - cudnn
Create ⚠Tensor Transform Descriptor - Create an empty tensor transform descriptor
- cudnn
DeriveBN ⚠Tensor Descriptor - cudnn
Derive ⚠Norm Tensor Descriptor - cudnn
Destroy ⚠ - cudnn
Destroy ⚠Activation Descriptor - cudnn
Destroy ⚠Algorithm Descriptor - cudnn
Destroy ⚠Algorithm Performance - cudnn
Destroy ⚠Attn Descriptor - cudnn
DestroyCTC ⚠Loss Descriptor - cudnn
Destroy ⚠Convolution Descriptor - cudnn
Destroy ⚠Dropout Descriptor - cudnn
Destroy ⚠Filter Descriptor - cudnn
Destroy ⚠Fused OpsConst Param Pack - cudnn
Destroy ⚠Fused OpsPlan - cudnn
Destroy ⚠Fused OpsVariant Param Pack - cudnn
DestroyLRN ⚠Descriptor - cudnn
Destroy ⚠OpTensor Descriptor - cudnn
Destroy ⚠PersistentRNN Plan - cudnn
Destroy ⚠Pooling Descriptor - cudnn
DestroyRNN ⚠Data Descriptor - cudnn
DestroyRNN ⚠Descriptor - cudnn
Destroy ⚠Reduce Tensor Descriptor - cudnn
Destroy ⚠SeqData Descriptor - cudnn
Destroy ⚠Spatial Transformer Descriptor - cudnn
Destroy ⚠Tensor Descriptor - cudnn
Destroy ⚠Tensor Transform Descriptor - Destroys a previously created tensor transform descriptor.
- cudnn
Divisive ⚠Normalization Backward - cudnn
Divisive ⚠Normalization Forward - cudnn
Dropout ⚠Backward - cudnn
Dropout ⚠Forward - cudnn
Dropout ⚠GetReserve Space Size - cudnn
Dropout ⚠GetStates Size - cudnn
Find ⚠Convolution Backward Data Algorithm - cudnn
Find ⚠Convolution Backward Data Algorithm Ex - cudnn
Find ⚠Convolution Backward Filter Algorithm - cudnn
Find ⚠Convolution Backward Filter Algorithm Ex - cudnn
Find ⚠Convolution Forward Algorithm - cudnn
Find ⚠Convolution Forward Algorithm Ex - cudnn
FindRNN ⚠Backward Data Algorithm Ex - cudnn
FindRNN ⚠Backward Weights Algorithm Ex - cudnn
FindRNN ⚠Forward Inference Algorithm Ex - cudnn
FindRNN ⚠Forward Training Algorithm Ex - cudnn
Fused ⚠OpsExecute - cudnn
GetActivation ⚠Descriptor - cudnn
GetActivation ⚠Descriptor Swish Beta - cudnn
GetAlgorithm ⚠Descriptor - cudnn
GetAlgorithm ⚠Performance - cudnn
GetAlgorithm ⚠Space Size - cudnn
GetAttn ⚠Descriptor - cudnn
GetBatch ⚠Normalization Backward ExWorkspace Size - cudnn
GetBatch ⚠Normalization Forward Training ExWorkspace Size - cudnn
GetBatch ⚠Normalization Training ExReserve Space Size - cudnn
GetCTC ⚠Loss Descriptor - cudnn
GetCTC ⚠Loss Descriptor Ex - cudnn
GetCTC ⚠Loss Descriptor_ v8 - cudnn
GetCTC ⚠Loss Workspace Size - cudnn
GetCTC ⚠Loss Workspace Size_ v8 - cudnn
GetCallback ⚠ - cudnn
GetConvolution2d ⚠Descriptor - cudnn
GetConvolution2d ⚠Forward Output Dim - cudnn
GetConvolution ⚠Backward Data Algorithm MaxCount - cudnn
GetConvolution ⚠Backward Data Algorithm_ v7 - cudnn
GetConvolution ⚠Backward Data Workspace Size - cudnn
GetConvolution ⚠Backward Filter Algorithm MaxCount - cudnn
GetConvolution ⚠Backward Filter Algorithm_ v7 - cudnn
GetConvolution ⚠Backward Filter Workspace Size - cudnn
GetConvolution ⚠Forward Algorithm MaxCount - cudnn
GetConvolution ⚠Forward Algorithm_ v7 - cudnn
GetConvolution ⚠Forward Workspace Size - cudnn
GetConvolution ⚠Group Count - cudnn
GetConvolution ⚠Math Type - cudnn
GetConvolution ⚠NdDescriptor - cudnn
GetConvolution ⚠NdForward Output Dim - cudnn
GetConvolution ⚠Reorder Type - cudnn
GetCudart ⚠Version - cudnn
GetDropout ⚠Descriptor - cudnn
GetError ⚠String - cudnn
GetFilter4d ⚠Descriptor - cudnn
GetFilter ⚠NdDescriptor - cudnn
GetFilter ⚠Size InBytes - cudnn
GetFolded ⚠Conv Backward Data Descriptors - cudnn
GetFused ⚠OpsConst Param Pack Attribute - cudnn
GetFused ⚠OpsVariant Param Pack Attribute - cudnn
GetLRN ⚠Descriptor - cudnn
GetMulti ⚠Head Attn Buffers - cudnn
GetMulti ⚠Head Attn Weights - cudnn
GetNormalization ⚠Backward Workspace Size - cudnn
GetNormalization ⚠Forward Training Workspace Size - cudnn
GetNormalization ⚠Training Reserve Space Size - cudnn
GetOp ⚠Tensor Descriptor - cudnn
GetPooling2d ⚠Descriptor - cudnn
GetPooling2d ⚠Forward Output Dim - cudnn
GetPooling ⚠NdDescriptor - cudnn
GetPooling ⚠NdForward Output Dim - cudnn
GetProperty ⚠ - cudnn
GetRNN ⚠Backward Data Algorithm MaxCount - cudnn
GetRNN ⚠Backward Weights Algorithm MaxCount - cudnn
GetRNN ⚠Bias Mode - cudnn
GetRNN ⚠Data Descriptor - cudnn
GetRNN ⚠Descriptor_ v6 - cudnn
GetRNN ⚠Descriptor_ v8 - cudnn
GetRNN ⚠Forward Inference Algorithm MaxCount - cudnn
GetRNN ⚠Forward Training Algorithm MaxCount - cudnn
GetRNN ⚠LinLayer Bias Params - cudnn
GetRNN ⚠LinLayer Matrix Params - cudnn
GetRNN ⚠Matrix Math Type - cudnn
GetRNN ⚠Padding Mode - cudnn
GetRNN ⚠Params Size - cudnn
GetRNN ⚠Projection Layers - cudnn
GetRNN ⚠Temp Space Sizes - cudnn
GetRNN ⚠Training Reserve Size - cudnn
GetRNN ⚠Weight Params - cudnn
GetRNN ⚠Weight Space Size - cudnn
GetRNN ⚠Workspace Size - cudnn
GetReduce ⚠Tensor Descriptor - cudnn
GetReduction ⚠Indices Size - cudnn
GetReduction ⚠Workspace Size - cudnn
GetSeq ⚠Data Descriptor - cudnn
GetStream ⚠ - cudnn
GetTensor4d ⚠Descriptor - cudnn
GetTensor ⚠NdDescriptor - cudnn
GetTensor ⚠Size InBytes - cudnn
GetTensor ⚠Transform Descriptor - Retrieves the values stored in a previously initialized tensor transform descriptor.
- cudnn
GetVersion ⚠ - cudnn
Im2Col ⚠ - cudnn
Init ⚠Transform Dest - Create a destination descriptor for cudnnTransformTensor
- cudnnLRN
Cross ⚠Channel Backward - cudnnLRN
Cross ⚠Channel Forward - cudnn
Make ⚠Fused OpsPlan - cudnn
Multi ⚠Head Attn Backward Data - cudnn
Multi ⚠Head Attn Backward Weights - cudnn
Multi ⚠Head Attn Forward - cudnn
Normalization ⚠Backward - cudnn
Normalization ⚠Forward Inference - cudnn
Normalization ⚠Forward Training - cudnn
OpTensor ⚠ - cudnn
OpsInfer ⚠Version Check - cudnn
OpsTrain ⚠Version Check - cudnn
Pooling ⚠Backward - cudnn
Pooling ⚠Forward - cudnn
Query ⚠Runtime Error - cudnnRNN
Backward ⚠Data - cudnnRNN
Backward ⚠Data Ex - cudnnRNN
Backward ⚠Data_ v8 - cudnnRNN
Backward ⚠Weights - cudnnRNN
Backward ⚠Weights Ex - cudnnRNN
Backward ⚠Weights_ v8 - cudnnRNN
Forward ⚠ - cudnnRNN
Forward ⚠Inference - cudnnRNN
Forward ⚠Inference Ex - cudnnRNN
Forward ⚠Training - cudnnRNN
Forward ⚠Training Ex - cudnnRNN
GetClip ⚠ - cudnnRNN
GetClip_ ⚠v8 - cudnnRNN
SetClip ⚠ - cudnnRNN
SetClip_ ⚠v8 - cudnn
Reduce ⚠Tensor - cudnn
Reorder ⚠Filter AndBias - cudnn
Restore ⚠Algorithm - cudnn
Restore ⚠Dropout Descriptor - cudnn
Save ⚠Algorithm - cudnn
Scale ⚠Tensor - cudnn
SetActivation ⚠Descriptor - cudnn
SetActivation ⚠Descriptor Swish Beta - cudnn
SetAlgorithm ⚠Descriptor - cudnn
SetAlgorithm ⚠Performance - cudnn
SetAttn ⚠Descriptor - cudnn
SetCTC ⚠Loss Descriptor - cudnn
SetCTC ⚠Loss Descriptor Ex - cudnn
SetCTC ⚠Loss Descriptor_ v8 - cudnn
SetCallback ⚠ - cudnn
SetConvolution2d ⚠Descriptor - cudnn
SetConvolution ⚠Group Count - cudnn
SetConvolution ⚠Math Type - cudnn
SetConvolution ⚠NdDescriptor - cudnn
SetConvolution ⚠Reorder Type - cudnn
SetDropout ⚠Descriptor - cudnn
SetFilter4d ⚠Descriptor - cudnn
SetFilter ⚠NdDescriptor - cudnn
SetFused ⚠OpsConst Param Pack Attribute - cudnn
SetFused ⚠OpsVariant Param Pack Attribute - cudnn
SetLRN ⚠Descriptor - cudnn
SetOp ⚠Tensor Descriptor - cudnn
SetPersistentRNN ⚠Plan - cudnn
SetPooling2d ⚠Descriptor - cudnn
SetPooling ⚠NdDescriptor - cudnn
SetRNN ⚠Algorithm Descriptor - cudnn
SetRNN ⚠Bias Mode - cudnn
SetRNN ⚠Data Descriptor - cudnn
SetRNN ⚠Descriptor_ v6 - cudnn
SetRNN ⚠Descriptor_ v8 - cudnn
SetRNN ⚠Matrix Math Type - cudnn
SetRNN ⚠Padding Mode - cudnn
SetRNN ⚠Projection Layers - cudnn
SetReduce ⚠Tensor Descriptor - cudnn
SetSeq ⚠Data Descriptor - cudnn
SetSpatial ⚠Transformer NdDescriptor - cudnn
SetStream ⚠ - cudnn
SetTensor ⚠ - cudnn
SetTensor4d ⚠Descriptor - cudnn
SetTensor4d ⚠Descriptor Ex - cudnn
SetTensor ⚠NdDescriptor - cudnn
SetTensor ⚠NdDescriptor Ex - cudnn
SetTensor ⚠Transform Descriptor - Initialize a previously created tensor transform descriptor.
- cudnn
Softmax ⚠Backward - cudnn
Softmax ⚠Forward - cudnn
Spatial ⚠TfGrid Generator Backward - cudnn
Spatial ⚠TfGrid Generator Forward - cudnn
Spatial ⚠TfSampler Backward - cudnn
Spatial ⚠TfSampler Forward - cudnn
Transform ⚠Filter - cudnn
Transform ⚠Tensor - cudnn
Transform ⚠Tensor Ex - tensor_
vec_ id_ c - Return C Handle for a Vector of Tensor Descriptors
Type Aliases§
- CUuuid
- __
int32_ t - __
int64_ t - __
uint32_ t - cuda
Array_ const_ t - CUDA array (as source copy argument)
- cuda
Array_ t - CUDA array
- cuda
Event_ t - CUDA event types
- cuda
External Memory_ t - CUDA external memory
- cuda
External Semaphore_ t - CUDA external semaphore
- cuda
Function_ t - CUDA function
- cuda
Graph Exec_ t - CUDA executable (launchable) graph
- cuda
Graph Node_ t - CUDA graph node.
- cuda
Graph_ t - CUDA graph
- cuda
Graphics Resource_ t - CUDA graphics resource types
- cuda
Host Fn_ t - CUDA host function \param userData Argument value passed to the function
- cuda
IpcEvent Handle_ t - CUDA IPC event handle
- cuda
IpcMem Handle_ t - CUDA IPC memory handle
- cuda
MemPool_ t - CUDA memory pool
- cuda
Mipmapped Array_ const_ t - CUDA mipmapped array (as source argument)
- cuda
Mipmapped Array_ t - CUDA mipmapped array
- cuda
Stream Callback_ t - Type of stream callback functions. \param stream The stream as passed to ::cudaStreamAddCallback, may be NULL. \param status ::cudaSuccess or any persistent error on the stream. \param userData User parameter provided at registration.
- cuda
Stream_ t - CUDA stream
- cuda
Surface Object_ t - An opaque value that represents a CUDA Surface object
- cuda
Texture Object_ t - An opaque value that represents a CUDA texture object
- cudaUUID_
t - cuda
User Object_ t - CUDA user object for graphs
- cudnn
Activation Descriptor_ t - cudnn
Algorithm Descriptor_ t - cudnn
Algorithm Performance_ t - cudnn
Algorithm_ t - cudnn
Attn Descriptor_ t - cudnn
Attn Query Map_ t - cudnn
Backend Descriptor_ t - cudnnCTC
Loss Descriptor_ t - cudnn
Callback_ t - cudnn
Convolution BwdData Algo Perf_ t - cudnn
Convolution BwdFilter Algo Perf_ t - cudnn
Convolution Descriptor_ t - cudnn
Convolution FwdAlgo Perf_ t - cudnn
Debug_ t - cudnn
Dropout Descriptor_ t - cudnn
Filter Descriptor_ t - cudnn
Fused OpsConst Param Pack_ t - cudnn
Fused OpsPlan_ t - cudnn
Fused OpsVariant Param Pack_ t - cudnn
Handle_ t - cudnnLRN
Descriptor_ t - cudnn
OpTensor Descriptor_ t - cudnn
PersistentRNN Plan_ t - cudnn
Pooling Descriptor_ t - cudnnRNN
Data Descriptor_ t - cudnnRNN
Descriptor_ t - cudnnRNN
Padding Mode_ t - cudnn
Reduce Tensor Descriptor_ t - cudnn
SeqData Descriptor_ t - cudnn
Spatial Transformer Descriptor_ t - cudnn
Tensor Descriptor_ t - cudnn
Tensor Transform Descriptor_ t - library
Property Type - library
Property Type_ t
Unions§
- cuda
External Memory Handle Desc__ bindgen_ ty_ 1 - cuda
External Semaphore Handle Desc__ bindgen_ ty_ 1 - cuda
External Semaphore Signal Params__ bindgen_ ty_ 1__ bindgen_ ty_ 2 - cuda
External Semaphore Wait Params__ bindgen_ ty_ 1__ bindgen_ ty_ 2 - cuda
Kernel Node Attr Value - Graph kernel node attributes union, used with ::cudaGraphKernelNodeSetAttribute/::cudaGraphKernelNodeGetAttribute
- cuda
Resource Desc__ bindgen_ ty_ 1 - cuda
Stream Attr Value - Stream attributes union used with ::cudaStreamSetAttribute/::cudaStreamGetAttribute
- cudnn
Algorithm Union Struct_ Algorithm