MATLAB: What is the difference between data types single and double in simulink

datatypesimulink

I tried to change the data type uint16 to double and i was getting an output of 10 digit number for inputs of a constant decimal number and 1 digit number.

But when I tried the same with data type single, I got the right anwer (with decimal points).

Kindly explain.

Best Answer

Your floating point data (Decimal Data) can be represented either via Single data type of Double data type. The basic difference being the precision between the two types. Now contary to the name, its not the precision that doubles. Its the number of Bits used in your computer memory for storing the data that doubles. Single datatype uses 32 bits of memory whereas Double uses 64 bits of memory. With higher precision, you get better accuracy but at the cost of memory and processing. This being a nutshell summary, you can google out for details. Hope this helps!

Related Solutions

MATLAB: Do I get an integer when I perform an operation on an integer and a double

This behavior is expected. There are two general reasons why returning an integer is preferable in this situation: First, integer data types are often used in situations where limiting memory usage is important, so it is useful that MATLAB returns a result that similarly conserves memory. Secondly, doubles are the default data type in MATLAB, so in a simple command such as:

>> int8(4)+1

the number 1 is a double, but the user would reasonably expect the output to be an int8. Making the output of an operation between doubles and integers always be an integer ensures this.

It is important to note that, while doubles and integers can be combined using an operator such as addition, subtraction, etc., integers of different classes (such as int8 and int16) cannot be combined using an operator.

MATLAB: Is there a global setting that makes Stateflow use single precision floats

The ability to treat all numbers and variables to be of type single by default is not available in Stateflow.

Stateflow treats all numeric constants with decimals as doubles. This is in accordance to the standard C behavior as Stateflow follows C syntax. Hence, both LHS and RHS in the expression should be of type single to avoid type casting. Hence, if we want to treat a decimal value like 3.0 (which is double by default) to be single, it can be done by adding F after the number (eg 0.3F will be considered as single while 0.3 will be taken as double by default). When we add two numbers, one of which is single and other a double, a singlewill upcast itself to be of type double.

Another way is to typecast each number to be of datatype 'single' within flowchart. eg:- single(0.3) will treat the numer 0.3 as single.

Variables can be defined as 'single' in the parameter settings if we want to treat numbers assigned to those variables as single. It can be done as follows:-

1. Open Model Explorer

2. Select the data inside the "Contents" pane

3. Modifying the "DataType" field of one of them

Attached is a test model to demonstrate the same where the numbers 2 and 0.5 are added. Here 0.5 is typecasted as single thereby adding two numbers of type single. On generating the report and looking at testmodel.c file, expressions like typecasts int_T and int32_T during addition of two numbers are no longer visible.

This is the default behaviour of Simulink during simulation and code generation and the functionality to make all the numbers and variables as single by default is currently not available.

Best Answer

Related Solutions

MATLAB: Do I get an integer when I perform an operation on an integer and a double

MATLAB: Is there a global setting that makes Stateflow use single precision floats

Related Question