On Sun, 20 Feb 2005 20:34:14 -0800
Robert Roessler <roessler@rftp.com> wrote:

> Actually, emulating the VC7 "/QIfist" does NOT cause "all int_to_float 
> to behave like a round instead of a truncate" - it does exactly what 
> we are already talking about: do the int_of_float with a "bare" 
> FIST[P], operating in whatever rounding mode the hardware is already 
> in (presumably one we want and expect it to be in).

Implementing your proposal, means that int_of_float will change
behaviour depending on a compiler flag. I think thats a bad
idea.

My proposal leaves int_of_float just as it is and defines a new
function:

    val round_to_int : float -> int

which expand to a FISTPL instruction and one or two glue 
instructions.

Erik
-- 
+-----------------------------------------------------------+
  Erik de Castro Lopo  nospam@mega-nerd.com (Yes it's valid)
+-----------------------------------------------------------+
Good advice for everyone : stay away from churches, mosques and
synagouges.