ReduceXorSum

Applicability

Product

Supported

Atlas 350 Accelerator Card

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 AI Core

Atlas inference product Vector Core

x

Atlas training product

x

Function Usage

Performs the XOR (bitwise XOR) operation by element and computes the sum of the results using ReduceSum.

If the final calculation result exceeds the int16 range [-32768, 32767], the output will be -32768 or 32767.

Prototype

  • Pass the temporary space through the sharedTmpBuffer input parameter.
    1
    2
    template <typename T, bool isReuseSource = false>
    __aicore__ inline void ReduceXorSum(LocalTensor<T>& dstTensor, const LocalTensor<T>& src0Tensor, const LocalTensor<T>& src1Tensor, LocalTensor<uint8_t>& sharedTmpBuffer, const uint32_t calCount)
    
  • Allocate the temporary space through the API framework.
    1
    2
    template <typename T, bool isReuseSource = false>
    __aicore__ inline void ReduceXorSum(LocalTensor<T>& dstTensor, const LocalTensor<T>& src0Tensor, const LocalTensor<T>& src1Tensor, const uint32_t calCount);
    

Due to the internal implementation of this API, which requires the storage of XOR outcomes and the execution of other operations, additional temporary space is required to store intermediate variables generated during computation. The temporary space can be passed by developers through the sharedTmpBuffer input parameter or allocated through the API framework.

  • When the sharedTmpBuffer input parameter is used for passing the temporary space, the tensor serves as the temporary space. In this case, the API framework is not required for temporary space allocation. This enables developers to manage the sharedTmpBuffer space and reuse the buffer after calling the API, so that the buffer is not repeatedly allocated or deallocated, improving the flexibility and buffer utilization.
  • When the API framework is used for temporary space allocation, developers do not need to allocate the space, but must reserve the required size for the temporary space.

If sharedTmpBuffer is used, developers must allocate space for the tensor. If the API framework is used, developers must reserve the temporary space. To obtain the size of the temporary space (BufferSize) to be reserved, use the API provided in GetReduceXorSumMaxMinTmpSize.

Parameters

Table 1 Template parameters

Parameter

Description

T

Data type of the operand.

For the Atlas 350 Accelerator Card, the supported data type is int16_t.

For the Atlas A3 training product/Atlas A3 inference product, the supported data type is int16_t.

For the Atlas A2 training product/Atlas A2 inference product, the supported data type is int16_t.

For the Atlas inference product AI Core, the supported data type is int16_t.

isReuseSource

Whether the source operand can be modified. The default value is false. If developers allow the source operand to be modified, enable this parameter, to reduce memory space usage.

If this parameter is set to true, the src0Tensor and src1Tensor memory space is reused during internal computation of this API to reduce memory space usage. If this parameter is set to false, the src0Tensor and src1Tensor memory space is not reused during internal computation of this API.

For details about how to use isReuseSource, see Example 4.

Table 2 API parameters

Parameter

Input/Output

Description

dstTensor

Output

Destination operand.

The type is LocalTensor, and TPosition can be VECIN, VECCALC, or VECOUT.

The output value needs to be saved in a space with a size of sizeof(T). Developers need to allocate the actual buffer space to dstTensor based on this size and the framework's alignment requirements.

NOTE:

The size of allocated buffer must be 32-byte aligned according to the framework's requirements. If the value of sizeof(T) is not 32-byte aligned, it should be rounded up to the nearest multiple of 32 bytes. The extra buffer space allocated for alignment purposes should not be filled with values, but rather left with random values.

src0Tensor

Input

Source operand 0.

The type is LocalTensor, and TPosition can be VECIN, VECCALC, or VECOUT.

The source operand must have the same data type as the destination operand.

src1Tensor

Input

Source operand 1.

The type is LocalTensor, and TPosition can be VECIN, VECCALC, or VECOUT.

The source operand must have the same data type as the destination operand.

sharedTmpBuffer

Input

Temporary buffer.

The type is LocalTensor, and TPosition can be VECIN, VECCALC, or VECOUT.

This parameter is used to store intermediate variables during ReduceXorSum computation and is provided by developers.

For details about how to obtain the temporary space size (BufferSize), see GetReduceXorSumMaxMinTmpSize.

calCount

Input

Number of elements involved in the computation.

Returns

None

Constraints

  • The source operand address must not overlap the destination operand address.
  • The address of sharedTmpBuffer cannot overlap that of the source or destination operand.
  • Ensure that calCount is less than or equal to the element range of src0Tensor and src1Tensor.
  • If the final calculation result exceeds the int16 range [-32768, 32767], the output will be -32768 or 32767.
  • For the Atlas inference product AI Core, the intermediate computation data is stored in half type. The final result error is larger compared with other processors.

Examples

1
2
3
4
// Template parameter: The operand data type is int16. The value false indicates that the source operand cannot be modified.
// dstLocal: tensor of the output data; src0Local: source operand 0; src1Local: source operand 1
// sharedTmpBuffer: temporary buffer, with 32 elements involved in computation.
AscendC::ReduceXorSum<int16_t, false>(dstLocal, src0Local, src1Local, sharedTmpBuffer, 32);
Result example:
1
2
3
4
5
The input and output data type is int16_t.
Input (src0Local): [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]
Input (src1Local): [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1]

Output (dstLocal): [32 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0] // Only 32 is a valid value.