I am trying to understand floating point arithmetic a little better. Can somebody please help me understand what is going on in the below example?

using System;

public class FloatingPointMadness
{
    public static void Main(string[] args)
    {
        float f = 0.1f;
        Console.WriteLine ($"Real value is: {f:F16}");
    }
}

I am using .Net Framework 4.7.2. Since the decimal 0.1 cannot be represented exactly with floating point arithmetic, I would expect this to produce something like:

Real value is: 0.1000000014901161

But instead I get:

Real value is: 0.1000000000000000

Online example. If I change to using the .NET 9 compiler, I get what I expect. What is going on here?

I am trying to understand floating point arithmetic a little better. Can somebody please help me understand what is going on in the below example?

using System;

public class FloatingPointMadness
{
    public static void Main(string[] args)
    {
        float f = 0.1f;
        Console.WriteLine ($"Real value is: {f:F16}");
    }
}

I am using .Net Framework 4.7.2. Since the decimal 0.1 cannot be represented exactly with floating point arithmetic, I would expect this to produce something like:

Real value is: 0.1000000014901161

But instead I get:

Real value is: 0.1000000000000000

Online example. If I change to using the .NET 9 compiler, I get what I expect. What is going on here?

• c#
• formatting
• precision

• Both 0.1000000014901161 and 0.1000000000000000 are not real values. They are merely the string representing data, not the data itself. The data is certain bitmap representing a floating-point object, and it is not a mathematical real number, this is a floating-point number. – Sergey A Kryukov Commented Feb 6 at 17:14
• 1 Also, you should never 100% rely on the string representation of data, especially if this is floating point. Strings should only be used to look at the data using human eyes. Strictly speaking, if you really want to store floating-point data in a text file, for example, and if you need accuracy, you have to store the number bitmap, not human-readable format. It makes the number even less accurate than floating-point data itself. – Sergey A Kryukov Commented Feb 6 at 17:18
• See the detailed description at https://learn.microsoft.com/en-us/dotnet/csharp/language-reference/builtin-types/floating-point-numeric-types. – OldBoy Commented Feb 6 at 17:24
• I would expect Why would you expect that? – mjwills Commented Feb 6 at 21:48
• Note that you should use "R" formatting specifier to "roundtrip" float values, i.e. ensure that you get the exact same value when parsing the string. – JonasH Commented Feb 7 at 8:44

1 Answer 1

It seems there is a parser and formatter compliance between .Net framework and .NET (or .Net Core). .NET uses the standard IEEE 754-2008 it seems.

I've tried your code in .Net framework and .NET (from 3.1 onwards) it behaves as you mentioned.

The reason is already answered in the below stackoverflow question: Rounding issues .Net Core 3.1 vs. .Net Core 2.0/.Net Framework

Hope this helps!