Why does C# host application change long double precision of C++ dll?

[Ostrenko]

Hello all,

I developed dll in Intel C++ that applies long double arithmetic operations.
My dll allocates 128 bit for each long double variable and performs 19 digits precision operations if I use C++ or Delphi host application.
In case of C# host application the dll allocates 128 bit also but number of significant digits is reduced to 15 (same with double).
How does it possible and what do I have to do to increase a number of significant digits to 19?

Thanks.

[GaryTexmo]

I don't think C# has support for a 128-bit floating point type. I did some googling and found this...

http://www.codeguru.com/forum/showthread.php?t=390950

I think if you want support for this, you'll either have to write a class in C# to support it, or create a wrapper class in C++ that you can use in C#.