Consider the following scenario:
You use In-Memory OLTP in Microsoft SQL Server.
You create a natively compiled module, such as a natively compiled user-defined function, that uses EXP functions.
You refer to that natively compiled module in another natively compiled module.
In this scenario, when you run the second natively compiled module, you receive an error message that resembles the following:
Msg 41328, Level 16, State 0
A floating point operation has overflowed.
Microsoft has confirmed that this is a problem in the Microsoft products that are listed in the "Applies to" section.