GetReduceSumMaxMinTmpSize

Function Usage

To perform ReduceSum computation in the kernel, developers need to reserve or allocate the temporary space. This API is used to obtain the maximum and minimum sizes of the temporary space to be reserved or allocated on the host. Developers can select a proper size within this range as the tiling parameter and pass it to the kernel.

To ensure correct functions, the temporary space to be reserved or allocated cannot be less than the minimum temporary space.
Within the range from the minimum to the maximum, as the temporary space increases, the API compute performance in the kernel can be optimized to some extent. To achieve better performance, reserve or allocate the space based on the actual buffer usage. The maximum temporary space of this API is currently equal to the minimum temporary space.

Prototype

      
           void GetReduceSumMaxMinTmpSize(const ge::Shape& srcShape, const ge::DataType dataType, ReducePattern pattern, bool isSrcInnerPad, bool isReuseSource, uint32_t& maxValue, uint32_t& minValue)

Parameters

Table 1 API parameters

API

Input/Output

Description

srcShape

Input

Input shape information. The value of this parameter must be the same as that of srcShape in the ReduceSum API.

dataType

Input

Input data type, which is of the ge::DataType type. For details about the definition of this type, see DataType. Currently, the supported data types are consistent with the template parameter T of the ReduceSum API.

pattern

Input

Specifies the computation axis of ReduceSum. ReducePattern type. This type is defined as follows, including the Reduced axis and Normal axis. pattern is a string composed of letters A (standing for Normal axis) and R (standing for Reduced axis), with the number of letters equal to the number of dimensions in the input vector. The value of this parameter must be the same as that of the pattern parameter of the ReduceSum API. Currently, only AscendC::ReducePattern::AR and AscendC::ReducePattern::RA are supported.

           
            
              
              
                enum class ReducePattern : uint32_t {
    AR = 0,
    RA = 1,
    R,
    ARA,
    ARAR,
    ARARA,
    ARARAR,
    ARARARA,
    ARARARAR,
    ARARARARA,
    RAR,
    RARA,
    RARAR,
    RARARA,
    RARARAR,
    RARARARA,
};

               

             

           
          

isSrcInnerPad

Input

Whether the innermost axis data to be computed is 32-byte aligned. The value of this parameter must be the same as that of the isSrcInnerPad parameter of the ReduceSum API.

isReuseSource

Input

Whether to reuse the space of the source operand input. The value of this parameter must be the same as that of the isReuseSource parameter of the ReduceSum API.

maxValue

Output

Maximum size of the temporary space required by ReduceSum computation. Any space exceeding this value will not be utilized by the API.

NOTE:

maxValue is for reference only and may be larger than the remaining space of the Unified Buffer. In this case, select a proper temporary space size based on the remaining space of the Unified Buffer.

minValue

Output

Minimum size of the temporary space required by ReduceSum computation. To ensure correct functions, the size of the temporary space to be reserved or allocated during API computation cannot be less than the value of this parameter.

Returns

None

Constraints

None

Examples

For a complete call example, see More Examples.

      
           // The input shape is 1024. The operator input data is of the float type. The source operand cannot be modified.
auto shape = ge::Shape({ 16, 32 });
uint32_t maxValue = 0;
uint32_t minValue = 0;
bool isSrcInnerPad = true;
bool isReuseSource = false;
AscendC::GetReduceSumMaxMinTmpSize(shape, ge::DataType::DT_FLOAT, AscendC::ReducePattern::AR, isSrcInnerPad, isReuseSource, maxValue, minValue);

Parent topic: ReduceSum APIs