GetWithOffset

Product Support

Product	Supported
Atlas A3 training products/Atlas A3 inference products	√
Atlas A2 training products/Atlas A2 inference products	√
Atlas 200I/500 A2 inference products	√
Atlas inference product's AI Core	√
Atlas inference product's Vector Core	x
Atlas training products	√

Function

Starts from the base address TBuf, offsets by a specified length, uses the offset address as the starting address, and extracts a tensor of the specified length.

Prototype

template <typename T>
__aicore__ inline LocalTensor<T> GetWithOffset(uint32_t size, uint32_t bufOffset)

Parameters

**Table 1** Template parameters
Parameter	Meaning
T	Data type of the tensor to be obtained.

**Table 2** Parameters
Parameter	Input/Output	Meaning
size	Input	Number of tensor elements to be obtained.
bufOffset	Input	Offset length from the start position, in bytes. The value must be 32-byte aligned.

Restrictions

The value of size is the number of elements in the tensor. The value of size x sizeof(T) + bufOffset cannot exceed the length during TBuf initialization.

The value of bufOffset must be 32-byte aligned.

Returns

Obtained LocalTensor.

Example

// Allocate buffer for TBuf initialization. The length of the allocated buffer is 1024 bytes.
AscendC::TPipe pipe;
AscendC::TBuf<AscendC::TPosition::VECCALC> calcBuf; // The template parameter is the VECCALC type in TPosition.
uint32_t byteLen = 1024;
pipe.InitBuffer(calcBuf, byteLen);
// Obtain a tensor from calcBuf using an offset of 64 bytes. The tensor is the buffer size of 128 elements of the int32_t type, which is 512 bytes.
AscendC::LocalTensor<int32_t> tempTensor1 = calcBuf.GetWithOffset<int32_t>(128, 64);

Parent topic: TBuf