Main Content

Sum Block Reorders Inputs

Open Model

This example shows how the Sum block reorders inputs. If you use a - sign as the first operation, the block reorders the inputs, if possible, to use a + operation. For example, in the expression output = -a-b+c, the Sum block reorders the input so that output = c-a-b. To initialize the accumulator, the Sum block uses the first + input port.

The block avoids performing a unary minus operation on the first operand a because doing so can change the value of a for fixed-point data types. In that case, the output value differs from the result of accumulating the values for a , b , and c .

Both the constant inputs use int8 data types The Sum block also uses int8 for the accumulator and output data types and has Saturate on integer overflow turned on. The Sum block reorders the inputs to give the ideal result of 127.

  1. Reorders inputs from (-Input1 + Input2) to (Input2 - Input1).

  2. Initializes the accumulator by using the first + input port. Accumulator = int8(-1) = -1

  3. Continues to accumulate values. Accumulator = Accumulator - int8(-128) = 127

  4. Calculates the block output. Output = int8(127) = 127

If the Sum block does not reorder the inputs, then you get the nonideal result of 126.

  1. Initializes the accumulator by using the first input port. Accumulator = int8(-(-128)) = 127

  2. Because saturation is on, the initial value of the accumulator saturates at 127 and does not wrap.

  3. Continues to accumulate values. Accumulator = Accumulator + int8(-1) = 126

  4. Calculates the block output. Output = int8(126) = 126

To explicitly specify a unary minus operation for output = -a-b+c, you can use the Unary Minus block in the Math Operations library.