Crate singe_cutensor_sys 

Raw FFI bindings for cuTENSOR and cuTENSORMg.

Prefer the safe singe-cutensor crate unless direct NVIDIA ABI access is required.

Structs

CUstream_st

_IO_FILE

_IO_codecvt

_IO_marker

_IO_wide_data

__BindgenBitfieldUnit

cuTensorMpContractionDescriptor

cuTensorMpContractionPlan

cuTensorMpHandle

cuTensorMpOperationDescriptor

cuTensorMpPlan

cuTensorMpPlanPreference

cuTensorMpTensorDescriptor

cutensorBlockSparseTensorDescriptor

cutensorComputeDescriptor

cutensorHandle

cutensorMgContractionDescriptor_s

cutensorMgContractionFind_s

cutensorMgContractionPlan_s

cutensorMgCopyDescriptor_s

cutensorMgCopyPlan_s

cutensorMgHandle_s

cutensorMgTensorDescriptor_s

cutensorOperationDescriptor

cutensorPlan

cutensorPlanPreference

cutensorTensorDescriptor

Enums

cudaDataType_t

cutensorAlgo_t

Allows users to specify the algorithm to be used for performing the desired tensor operation.

cutensorAutotuneMode_t

This enum determines the mode with respect to cuTENSOR’s auto-tuning capability.

cutensorCacheMode_t

This enum defines what is considered a cache hit.

cutensorComputeType_t

cutensorDataType_t

cutensorJitMode_t

This enum determines the mode with respect to cuTENSOR’s just-in-time compilation capability.

cutensorMgAlgo_t

Represents the selected algorithm when planning for a contraction operation.

cutensorMgContractionFindAttribute_t

cutensorMgHostDevice_t

Enumerated device codes for host-side tensors.

cutensorMpAlgo_t

cutensorMpPlanAttribute_t

cutensorOperationDescriptorAttribute_t

This enum lists all attributes of a cutensorOperationDescriptor_t that can be modified (see cutensorOperationDescriptorSetAttribute and cutensorOperationDescriptorGetAttribute).

cutensorOperator_t

This enum captures all unary and binary element-wise operations supported by the cuTENSOR library.

cutensorPlanAttribute_t

This enum lists all attributes of a cutensorPlan_t object that can be retrieved via cutensorPlanGetAttribute.

cutensorPlanPreferenceAttribute_t

This enum lists all attributes of a cutensorPlanPreference_t object that can be modified.

cutensorStatus_t

cuTENSOR status type returns

cutensorWorksizePreference_t

This enum gives users finer control over the suggested workspace.

Constants

CUTENSORMP_MAJOR

CUTENSORMP_MINOR

CUTENSORMP_PATCH

CUTENSORMP_VERSION

CUTENSOR_MAJOR

CUTENSOR_MINOR

CUTENSOR_PATCH

CUTENSOR_VERSION

Statics

CUTENSOR_COMPUTE_DESC_3XTF32^⚠

CUTENSOR_COMPUTE_DESC_4X16F^⚠

CUTENSOR_COMPUTE_DESC_8XINT8^⚠

CUTENSOR_COMPUTE_DESC_9X16BF^⚠

CUTENSOR_COMPUTE_DESC_16BF^⚠

CUTENSOR_COMPUTE_DESC_16F^⚠

CUTENSOR_COMPUTE_DESC_32F^⚠

CUTENSOR_COMPUTE_DESC_64F^⚠

CUTENSOR_COMPUTE_DESC_TF32^⚠

CUTENSOR_C_4I^⚠

CUTENSOR_C_4U^⚠

CUTENSOR_C_8I^⚠

CUTENSOR_C_8U^⚠

CUTENSOR_C_16BF^⚠

CUTENSOR_C_16F^⚠

CUTENSOR_C_16I^⚠

CUTENSOR_C_16U^⚠

CUTENSOR_C_32F^⚠

CUTENSOR_C_32I^⚠

CUTENSOR_C_32U^⚠

CUTENSOR_C_64F^⚠

CUTENSOR_C_64I^⚠

CUTENSOR_C_64U^⚠

CUTENSOR_R_4I^⚠

CUTENSOR_R_4U^⚠

CUTENSOR_R_8I^⚠

CUTENSOR_R_8U^⚠

CUTENSOR_R_16BF^⚠

CUTENSOR_R_16F^⚠

CUTENSOR_R_16I^⚠

CUTENSOR_R_16U^⚠

CUTENSOR_R_32F^⚠

CUTENSOR_R_32I^⚠

CUTENSOR_R_32U^⚠

CUTENSOR_R_64F^⚠

CUTENSOR_R_64I^⚠

CUTENSOR_R_64U^⚠

Functions

cutensorBlockSparseContract^⚠

This routine computes the block-sparse tensor contraction $D = alpha \* A \* B + beta \* C$. $$ \mathcal{D}{{modes}\mathcal{D}} \gets \alpha \* \mathcal{A}{{modes}\mathcal{A}} B_{{modes}\mathcal{B}} + \beta \mathcal{C}{{modes}_\mathcal{C}} $$

cutensorContract^⚠

This routine computes the tensor contraction $D = alpha \* A \* B + beta \* C$. $$ \mathcal{D}{{modes}\mathcal{D}} \gets \alpha \* \mathcal{A}{{modes}\mathcal{A}} B_{{modes}\mathcal{B}} + \beta \mathcal{C}{{modes}_\mathcal{C}} $$

cutensorContractTrinary^⚠

This routine computes the tensor contraction $E = alpha \* A \* B \* C + beta \* D$. $$ \mathcal{E}{{modes}\mathcal{E}} \gets \alpha \* \mathcal{A}{{modes}\mathcal{A}} \mathcal{B}{{modes}\mathcal{B}} \mathcal{C}{{modes}\mathcal{C}} + \beta \mathcal{D}{{modes}\mathcal{D}} $$

cutensorCreate^⚠

Initializes the cuTENSOR library and allocates the memory for the library context.

cutensorCreateBlockSparseContraction^⚠

This function allocates a cutensorOperationDescriptor_t object that encodes a block-sparse tensor contraction of the form $D = \alpha \mathcal{A} \mathcal{B} + \beta \mathcal{C}$.

cutensorCreateBlockSparseTensorDescriptor^⚠

Create a block-sparse tensor descriptor.

cutensorCreateContraction^⚠

This function allocates a cutensorOperationDescriptor_t object that encodes a tensor contraction of the form $D = \alpha \mathcal{A} \mathcal{B} + \beta \mathcal{C}$.

cutensorCreateContractionTrinary^⚠

This function allocates a cutensorOperationDescriptor_t object that encodes a tensor contraction of the form $\mathcal{E} = \alpha \mathcal{A} \mathcal{B} \mathcal{C} + \beta \mathcal{D}$.

cutensorCreateElementwiseBinary^⚠

This function creates an operation descriptor for an elementwise binary operation.

cutensorCreateElementwiseTrinary^⚠

This function creates an operation descriptor that encodes an elementwise trinary operation.

cutensorCreatePermutation^⚠

This function creates an operation descriptor for a tensor permutation.

cutensorCreatePlan^⚠

This function allocates a cutensorPlan_t object, selects an appropriate kernel for a given operation (encoded by desc) and prepares a plan that encodes the execution.

cutensorCreatePlanPreference^⚠

Allocates the cutensorPlanPreference_t, enabling users to limit the applicable kernels for a given plan/operation.

cutensorCreateReduction^⚠

Creates a cutensorOperatorDescriptor_t object that encodes a tensor reduction of the form $D = alpha \* opReduce(opA(A)) + beta \* opC(C)$.

cutensorCreateTensorDescriptor^⚠

Creates a tensor descriptor.

cutensorDestroy^⚠

Frees all resources related to the provided library handle.

cutensorDestroyBlockSparseTensorDescriptor^⚠

Frees all resources related to the provided block-sparse tensor descriptor.

cutensorDestroyOperationDescriptor^⚠

Frees all resources related to the provided descriptor.

cutensorDestroyPlan^⚠

Frees all resources related to the provided plan.

cutensorDestroyPlanPreference^⚠

Frees all resources related to the provided preference.

cutensorDestroyTensorDescriptor^⚠

Frees all resources related to the provided tensor descriptor.

cutensorElementwiseBinaryExecute^⚠

Performs an element-wise tensor operation for two input tensors (see cutensorCreateElementwiseBinary)

cutensorElementwiseTrinaryExecute^⚠

Performs an element-wise tensor operation for three input tensors (see cutensorCreateElementwiseTrinary)

cutensorEstimateWorkspaceSize^⚠

Determines the required workspaceSize for the given operation encoded by desc.

cutensorGetCudartVersion^⚠

Returns version number of the CUDA runtime that cuTENSOR was compiled against.

cutensorGetErrorString^⚠

Returns the description string for an error code.

cutensorGetVersion^⚠

Returns Version number of the CUTENSOR library.

cutensorHandleReadPlanCacheFromFile^⚠

Reads a Plan-Cache from file and overwrites the cachelines of the provided handle.

cutensorHandleResizePlanCache^⚠

Resizes the plan cache.

cutensorHandleWritePlanCacheToFile^⚠

Writes the Plan-Cache (that belongs to the provided handle) to file.

cutensorLoggerForceDisable^⚠

This function disables logging for the entire run.

cutensorLoggerOpenFile^⚠

This function opens a logging output file in the given path.

cutensorLoggerSetCallback^⚠

This function sets the logging callback routine.

cutensorLoggerSetFile^⚠

This function sets the logging output file.

cutensorLoggerSetLevel^⚠

This function sets the value of the logging level.

cutensorLoggerSetMask^⚠

This function sets the value of the log mask.

cutensorMgContraction^⚠

Execute a contraction operation.

cutensorMgContractionFindSetAttribute^⚠

cutensorMgContractionGetWorkspace^⚠

Computes the workspace that is needed for the contraction.

cutensorMgCopy^⚠

Execute a copy operation.

cutensorMgCopyGetWorkspace^⚠

Computes the workspace that is needed for the copy.

cutensorMgCreate^⚠

Create a library handle.

cutensorMgCreateContractionDescriptor^⚠

Create a contraction descriptor.

cutensorMgCreateContractionFind^⚠

Create a contraction find.

cutensorMgCreateContractionPlan^⚠

Create a contraction plan.

cutensorMgCreateCopyDescriptor^⚠

Create a copy descriptor.

cutensorMgCreateCopyPlan^⚠

Create a copy plan.

cutensorMgCreateTensorDescriptor^⚠

Create a tensor descriptor.

cutensorMgDestroy^⚠

Destroy a library handle.

cutensorMgDestroyContractionDescriptor^⚠

Destroy a contraction descriptor.

cutensorMgDestroyContractionFind^⚠

Destroy a contraction find.

cutensorMgDestroyContractionPlan^⚠

Destroy a contraction plan.

cutensorMgDestroyCopyDescriptor^⚠

Destroy a copy descriptor and free all its previously-allocated resources.

cutensorMgDestroyCopyPlan^⚠

Destroy a copy plan.

cutensorMgDestroyTensorDescriptor^⚠

Destroy a tensor descriptor.

cutensorMpContract^⚠

Performs a distributed tensor contraction across multiple processes.

cutensorMpCreate^⚠

Initializes the cutensorMp library and creates a handle for distributed tensor operations.

cutensorMpCreateContraction^⚠

Creates an operation descriptor that encodes a distributed tensor contraction.

cutensorMpCreatePlan^⚠

Creates an execution plan for distributed tensor contractions.

cutensorMpCreatePlanPreference^⚠

Creates a plan preference object for controlling distributed tensor operation planning.

cutensorMpCreateTensorDescriptor^⚠

Creates a distributed tensor descriptor for multi-process tensor operations.

cutensorMpDestroy^⚠

Frees all resources associated with the provided cutensorMp handle.

cutensorMpDestroyOperationDescriptor^⚠

Frees all resources related to the provided distributed contraction descriptor.

cutensorMpDestroyPlan^⚠

Frees all resources related to the provided distributed contraction plan.

cutensorMpDestroyPlanPreference^⚠

Frees all resources related to the provided plan preference object.

cutensorMpDestroyTensorDescriptor^⚠

Frees all resources related to the provided distributed tensor descriptor.

cutensorMpPlanGetAttribute^⚠

Retrieves information about an already-created plan (see cutensorPlanAttribute_t).

cutensorOperationDescriptorGetAttribute^⚠

This function retrieves an attribute of the provided cutensorOperationDescriptor_t object (see cutensorOperationDescriptorAttribute_t).

cutensorOperationDescriptorSetAttribute^⚠

Set attribute of a cutensorOperationDescriptor_t object.

cutensorPermute^⚠

Performs the tensor permutation that is encoded by plan (see cutensorCreatePermutation).

cutensorPlanGetAttribute^⚠

Retrieves information about an already-created plan (see cutensorPlanAttribute_t)

cutensorPlanPreferenceGetAttribute^⚠

Get attribute of a cutensorPlanPreference_t object.

cutensorPlanPreferenceSetAttribute^⚠

Set attribute of a cutensorPlanPreference_t object.

cutensorReadKernelCacheFromFile^⚠

Reads a kernel cache from file and adds all non-existing JIT compiled kernels to the kernel cache.

cutensorReduce^⚠

Performs the tensor reduction that is encoded by plan (see cutensorCreateReduction).

cutensorWriteKernelCacheToFile^⚠

Writes the —per library— kernel cache to file.

Type Aliases

FILE

_IO_lock_t

__off64_t

__off_t

__uint64_t

cutensorBlockSparseTensorDescriptor_t

Opaque structure representing a block-sparse tensor descriptor.

cutensorComputeDescriptor_t

Opaque structure representing a compute descriptor.

cutensorHandle_t

Opaque structure holding cuTENSOR’s library context.

cutensorLoggerCallback_t

A function pointer type for logging.

cutensorMgContractionDescriptor_t

Describes the contraction of two tensors into a third tensor with an optional source.

cutensorMgContractionFind_t

Describes the algorithmic details of implementing a tensor contraction.

cutensorMgContractionPlan_t

Describes a specific way to implement a contraction operation.

cutensorMgCopyDescriptor_t

Describes the copy of a tensor from one data layout to another.

cutensorMgCopyPlan_t

Describes a specific way to implement the copy operation.

cutensorMgHandle_t

Encodes the devices that participate in operations.

cutensorMgTensorDescriptor_t

Represents a tensor that may be distributed.

cutensorMpContractionDescriptor_t

cutensorMpContractionPlan_t

cutensorMpHandle_t

cutensorMpOperationDescriptor_t

cutensorMpPlanPreference_t

cutensorMpPlan_t

cutensorMpTensorDescriptor_t

cutensorOperationDescriptor_t

Opaque structure representing any type of problem descriptor (e.g., contraction, reduction, elementwise).

cutensorPlanPreference_t

Opaque structure that narrows down the space of applicable algorithms/variants/kernels.

cutensorPlan_t

Opaque structure representing a plan (e.g, contraction, reduction, elementwise).

cutensorTensorDescriptor_t

Opaque structure representing a tensor descriptor.

size_t