Skip to main content
Retrocomputing

You are not logged in. Your edit will be placed in a queue until it is peer reviewed.

We welcome edits that make the post easier to understand and more valuable for readers. Because community members review edits, please try to make the post substantially better than how you found it, for example, by fixing grammar or adding additional resources and hyperlinks.

3
  • 1
    Right, I was talking about the historical context in which x87 was the only FPU on x86, so no performance hit from using it. Good point about 68881 being a very similar architecture. Commented Apr 19, 2019 at 8:12
  • If memory serves, the 68000-family Macintosh could process computations with the 80-bit extended-precision type were faster than computations using 64-bit double-precision type, whether or not an FPU was present. The problem with the 80-bit type is that ANSI broke the C language by making the representation of a floating-point value passed to a non-prototyped function depend upon its precision, rather than specifying that any floating-point value that is passed to a non-prototyped function without being with a certain intrinsic will be passed as double, and any value which is wrapped... Commented Nov 12, 2020 at 20:57
  • ...in an _LDBL(x) intrinsic will be passed as long double [the exact behavior of the intrinsic would be up to the compiler, but a typical implementation could have it be a function return a struct of implementation-reserved type containing a long double]. The only convenient way for implementations to be compatible with code that expected to receive 64-bit double values without breaking the language was to interpret both double and long double as synonyms for the same 64-bit type. Commented Nov 12, 2020 at 21:00