correct round of reals?

[Uwe Zessin]

In article <XGUS4.3839$wYl.204453376 at newsb.telia.net>,
  "Fredrik Lundh" <effbot at telia.com> wrote:
> Peter Schneider-Kamp <petersc at stud.ntnu.no> wrote:
> > Thanks. Do you know if e.g. VC++ uses IEEE doubles?
>
> most modern platforms use IEEE, but I'm sure there are
> exceptions (OpenVMS?)

The OpenVMS operating system runs on 2 hardware platforms:

1) the VAX architecture (32-bit CISC - first model sold in 1978)
   it has several floating point formats.
   IIRC the IEEE definition came later and it somehow had to be
   different (the same reason why the DEC multinational character
   set is different than ISO Latin-1 :-(

2) the Alpha architecture (64-bit RISC - first sold, IIRC, in 1991/2)
   it can do VAX floating point (with some limitations, I beleive)
   and IEEE.
   Cameron Laird is correct that one can select the format with
   a compiler switch and the default is non-IEEE. The current
   Python port for OpenVMS Alpha uses just the compiler default.

--
Uwe Zessin


Sent via Deja.com http://www.deja.com/
Before you buy.