C++ API Reference for Intel® Data Analytics Acceleration Library 2018 Update 2
Abstract interface class for a data management component responsible for representation of data in the numeric format. This class declares the most general methods for data access. More...
Public Types | |
enum | MemoryStatus { notAllocated = 0, userAllocated = 1, internallyAllocated = 2 } |
Enumeration to specify the status of memory related to the Numeric Table. More... | |
enum | AllocationFlag { doNotAllocate = 0, notAllocate = 0, doAllocate = 1 } |
Enumeration to specify whether the Numeric Table must allocate memory. More... | |
enum AllocationFlag |
Enumerator | |
---|---|
doNotAllocate | Memory will not be allocated by Tensor |
notAllocate | Memory will not be allocated by Tensor
|
doAllocate | Memory will be allocated by Tensor when needed |
enum MemoryStatus |
Enumerator | |
---|---|
notAllocated | No memory allocated |
userAllocated | Memory allocated on user side |
internallyAllocated | Memory allocated and managed by Tensor |
|
pure virtual |
Allocates memory for a data set
Implemented in Tensor.
|
pure virtual |
Checks the correctness of this tensor
[in] | description | Additional information about error |
Implemented in Tensor.
|
pure virtual |
Deallocates the memory allocated for a data set
Implemented in Tensor.
|
pure virtual |
Returns new tensor with first dimension limited to one point
[in] | firstDimIndex | Index of the point in the first dimention |
Implemented in HomogenTensor< DataType >.
|
pure virtual |
Sets the number of dimensions in the Tensor
[in] | ndim | Number of dimensions |
[in] | dimSizes | Array with sizes for each dimension |
Implemented in HomogenTensor< DataType >.
|
pure virtual |
Sets the number and size of dimensions in the Tensor
[in] | dimensions | Collection with sizes for each dimension |
Implemented in HomogenTensor< DataType >.
For more complete information about compiler optimizations, see our Optimization Notice.