AllocTensor

Applicability

Product

Supported

Atlas A3 training product/Atlas A3 inference product

Atlas A2 training product/Atlas A2 inference product

Atlas 200I/500 A2 inference product

x

Atlas inference product

x

Atlas training product

x

Function Description

FlowBufferFactory is used to build shared memory to reduce input and output copying time.

AllocTensor allocates tensors based on the shape, data type, and alignment size. By default, the tensor is 64-byte aligned. You can specify the alignment size for performance tuning.

Prototype

1
std::shared_ptr<Tensor> AllocTensor(const std::vector<int64_t> &shape, TensorDataType dataType, uint32_t align = 512U)

Parameters

Parameter

Input/Output

Description

shape

Input

Shape of a tensor.

dataType

Input

Data type of a tensor.

align

Input

Alignment size for the allocated memory address. The value range is [32, 64, 128, 256, 512, 1024].

Returns

Pointer to the allocated tensor.

Exception Handling

NULL is returned if the tensor allocation fails.

Constraints

None