NVPL TENSOR Data Types¶
nvpltensorDataType_t
¶
-
enum nvpltensorDataType_t¶
This enum specifies the data precision. It is used when the data reference does not carry the type itself (e.g void *)
Values:
-
enumerator NVPLTENSOR_R_32F¶
real as a float
-
enumerator NVPLTENSOR_C_32F¶
complex as a pair of float numbers
-
enumerator NVPLTENSOR_R_64F¶
real as a double
-
enumerator NVPLTENSOR_C_64F¶
complex as a pair of double numbers
-
enumerator NVPLTENSOR_R_32F¶
nvpltensorComputeDescriptor_t
¶
-
typedef struct nvpltensorComputeDescriptor *nvpltensorComputeDescriptor_t¶
Opaque structure representing a tensor descriptor.
- NVPLTENSOR_EXTERN const nvpltensorComputeDescriptor_t NVPLTENSOR_COMPUTE_DESC_32F
floating-point: 8-bit exponent and 23-bit mantissa (aka float)
- NVPLTENSOR_EXTERN const nvpltensorComputeDescriptor_t NVPLTENSOR_COMPUTE_DESC_64F
floating-point: 11-bit exponent and 52-bit mantissa (aka double)
nvpltensorHandle_t
¶
-
typedef struct nvpltensorHandle *nvpltensorHandle_t¶
Opaque structure holding nvplTENSOR’s library context.
nvpltensorTensorDescriptor_t
¶
-
typedef struct nvpltensorTensorDescriptor *nvpltensorTensorDescriptor_t¶
Opaque structure representing a tensor descriptor.
nvpltensorOperationDescriptor_t
¶
-
typedef struct nvpltensorOperationDescriptor *nvpltensorOperationDescriptor_t¶
Opaque structure representing any type of problem descriptor (e.g., contraction, reduction, elementwise).
nvpltensorOperationDescriptorAttribute_t
¶
-
enum nvpltensorOperationDescriptorAttribute_t¶
This enum lists all attributes of a nvpltensorOperationDescriptor_t that can be modified (see nvpltensorOperationDescriptorSetAttribute and nvpltensorOperationDescriptorGetAttribute).
Values:
-
enumerator NVPLTENSOR_OPERATION_DESCRIPTOR_TAG¶
int32_t: enables users to distinguish two identical problems w.r.t. the sw-managed plan-cache. (default value: 0)
-
enumerator NVPLTENSOR_OPERATION_DESCRIPTOR_SCALAR_TYPE¶
nvpltensorDataType_t: data type of the scaling factors
-
enumerator NVPLTENSOR_OPERATION_DESCRIPTOR_FLOPS¶
float: number of floating-point operations necessary to perform this operation (assuming all scalar are not equal to zero, unless otherwise specified)
-
enumerator NVPLTENSOR_OPERATION_DESCRIPTOR_MOVED_BYTES¶
float: minimal number of bytes transferred from/to global-memory (assuming all scalar are not equal to zero, unless otherwise specified)
-
enumerator NVPLTENSOR_OPERATION_DESCRIPTOR_TAG¶
nvpltensorPlanPreference_t
¶
-
typedef struct nvpltensorPlanPreference *nvpltensorPlanPreference_t¶
Opaque structure that narrows down the space of applicable algorithms/variants/kernels.
nvpltensorPlanPreferenceAttribute_t
¶
-
enum nvpltensorPlanPreferenceAttribute_t¶
This enum lists all attributes of a nvpltensorPlanPreference_t object that can be modified.
Values:
-
enumerator NVPLTENSOR_PLAN_PREFERENCE_ALGO¶
nvpltensorAlgo_t: Fixes a certain nvpltensorAlgo_t
-
enumerator NVPLTENSOR_PLAN_PREFERENCE_ALGO¶
nvpltensorPlan_t
¶
-
typedef struct nvpltensorPlan *nvpltensorPlan_t¶
Opaque structure representing a plan (e.g, contraction, reduction, elementwise).
nvpltensorPlanAttribute_t
¶
-
enum nvpltensorPlanAttribute_t¶
This enum lists all attributes of a nvpltensorPlan_t object that can be retrieved via nvpltensorPlanGetAttribute.
Values:
-
enumerator NVPLTENSOR_PLAN_REQUIRED_WORKSPACE¶
uint64_t: exact required workspace in bytes that is needed to execute the plan
-
enumerator NVPLTENSOR_PLAN_REQUIRED_WORKSPACE¶
nvpltensorJitMode_t
¶
-
enum nvpltensorJitMode_t¶
This enum determines the mode w.r.t. nvplTENSOR’s just-in-time compilation capability.
Values:
-
enumerator NVPLTENSOR_JIT_MODE_NONE¶
Indicates that no kernel will be just-in-time compiled.
-
enumerator NVPLTENSOR_JIT_MODE_NONE¶
nvpltensorAlgo_t
¶
-
enum nvpltensorAlgo_t¶
Allows users to specify the algorithm to be used for performing the desired tensor operation.
Values:
-
enumerator NVPLTENSOR_ALGO_DEFAULT¶
A performance model chooses the appropriate algorithm and kernel.
-
enumerator NVPLTENSOR_ALGO_DEFAULT¶
nvpltensorWorksizePreference_t
¶
-
enum nvpltensorWorksizePreference_t¶
This enum gives users finer control over the suggested workspace.
This enum gives users finer control over the amount of workspace that is suggested by nvpltensorEstimateWorkspaceSize
Values:
-
enumerator NVPLTENSOR_WORKSPACE_MIN¶
Least memory requirement; at least one algorithm will be available.
-
enumerator NVPLTENSOR_WORKSPACE_DEFAULT¶
Aims to attain high performance while also reducing the workspace requirement.
-
enumerator NVPLTENSOR_WORKSPACE_MAX¶
Highest memory requirement; all algorithms will be available (choose this option if memory footprint is not a concern)
-
enumerator NVPLTENSOR_WORKSPACE_MIN¶
nvpltensorOperator_t
¶
-
enum nvpltensorOperator_t¶
This enum captures all unary and binary element-wise operations supported by the nvplTENSOR library.
Values:
-
enumerator NVPLTENSOR_OP_IDENTITY¶
Identity operator (i.e., elements are not changed)
-
enumerator NVPLTENSOR_OP_ADD¶
Addition of two elements.
-
enumerator NVPLTENSOR_OP_UNKNOWN¶
reserved for internal use only
-
enumerator NVPLTENSOR_OP_IDENTITY¶
nvpltensorStatus_t
¶
-
enum nvpltensorStatus_t¶
nvplTENSOR status type returns
The type is used for function status returns. All nvplTENSOR library functions return their status, which can have the following values.
Values:
-
enumerator NVPLTENSOR_STATUS_SUCCESS¶
The operation completed successfully.
-
enumerator NVPLTENSOR_STATUS_NOT_INITIALIZED¶
The opaque data structure was not initialized.
-
enumerator NVPLTENSOR_STATUS_ALLOC_FAILED¶
Resource allocation failed inside the nvplTENSOR library.
-
enumerator NVPLTENSOR_STATUS_INVALID_VALUE¶
An unsupported value or parameter was passed to the function (indicates an user error).
-
enumerator NVPLTENSOR_STATUS_INTERNAL_ERROR¶
An internal nvplTENSOR error has occurred.
-
enumerator NVPLTENSOR_STATUS_NOT_SUPPORTED¶
The requested operation is not supported.
-
enumerator NVPLTENSOR_STATUS_LICENSE_ERROR¶
The functionality requested requires some license and an error was detected when trying to check the current licensing.
-
enumerator NVPLTENSOR_STATUS_INSUFFICIENT_WORKSPACE¶
The provided workspace was insufficient.
-
enumerator NVPLTENSOR_STATUS_IO_ERROR¶
Indicates an error related to file I/O.
-
enumerator NVPLTENSOR_STATUS_SUCCESS¶
nvpltensorLoggerCallback_t
¶
-
typedef void (*nvpltensorLoggerCallback_t)(int32_t logLevel, const char *functionName, const char *message)¶
A function pointer type for logging.