MulOverflow

Function Usage

A template function to determine whether overflow occurs when two numbers are multiplied. If no overflow occurs, the correct computation result is returned.

Prototype

template<typename TLhs, typename TRhs, typename TRet>

bool MulOverflow(TLhs lhs, TRhs rhs, TRet &ret)

Parameters

**Table 1** Parameters in the template
Parameter	Description
TLhs	Data type of the left operand for the multiplication operation.
TRhs	Data type of the right operand for the multiplication operation.
TRet	Data type of the multiplication operation result.

**Table 2** Parameters
Parameter	Input/Output	Description
lhs	Input	Left operand of the multiplication operation.
rhs	Input	Right operand of the multiplication operation.
ret	Input	Multiplication result of the left and right operands. The result is valid only when the function return value is true.

Returns

The value true indicates that the computation fails and the return value of ret is invalid. The value false indicates that the computation is successful and the return value of ret is valid.

Constraints

None

Examples

ge::DataType out_data_type = ge::DT_FLOAT;
GE_ASSERT_GRAPH_SUCCESS(GetOutputDataType(context, out_data_type), "get data type failed");
GE_ASSERT_TRUE(out_data_type == ge::DataType::DT_INT32 || out_data_type == ge::DataType::DT_INT64,
               "only support DT_INT32 and DT_INT64, but out_data_type is %s",
               ge::TypeUtils::DataTypeToSerialString(out_data_type).c_str());
const auto is_malloc = (out_data_type == ge::DataType::DT_INT32);
const auto data_type_size = ge::GetSizeByDataType(out_data_type);
if (data_type_size <= 0) {
   // Error reported
}
size_t malloc_buffer_size = 0U;
if (ge::MulOverflow(static_cast<size_t>(data_type_size), Shape::kMaxDimNum, malloc_buffer_size)) {
   // Error reported
}
if (ge::AddOverflow(malloc_buffer_size, sizeof(GertTensorData), malloc_buffer_size)) {
   // Error reported
}

Parent topic: GE Namespace