Re: Why do some math functions return -0

[KHMan <keinhong@...>]

On 9/20/2010 12:45 AM, Mike McGonagle wrote:
> I am running on a Mac OS X 10.4.11 with lua 5.1.2, comiled in 2007.
>
> here are the first few lines of output...
>
> i:0.000000 r:0.000000 c:-0.000000 b:0.000000 br:-0.000000 bc:-0.000000
> i:1.000000 r:0.000000 c:1.000000 b:0.000000 br:-0.000000 bc:1.000000
> i:2.000000 r:0.000000 c:2.000000 b:0.000000 br:-0.000000 bc:2.000000
> i:3.000000 r:0.000000 c:3.000000 b:1.000000 br:-0.000000 bc:-0.000000
> i:4.000000 r:0.000000 c:4.000000 b:1.000000 br:-0.000000 bc:1.000000
> i:5.000000 r:0.000000 c:5.000000 b:1.000000 br:-0.000000 bc:2.000000
> i:6.000000 r:0.000000 c:6.000000 b:2.000000 br:-0.000000 bc:-0.000000
> i:7.000000 r:0.000000 c:7.000000 b:2.000000 br:-0.000000 bc:1.000000
> i:8.000000 r:0.000000 c:8.000000 b:2.000000 br:-0.000000 bc:2.000000
> i:9.000000 r:1.000000 c:-0.000000 b:0.000000 br:1.000000 bc:-0.000000
>
> As you notice, not all the 0's are negative. I don't know, maybe this
> really isn't an issue, as so far I have not run into anything other
> than this output being not exactly what I expect. I do know that I
> have a later version of lua on my machine, but I don't really have the
> time at the moment to check it... so maybe I should just try that and
> then get back to the list...

It looked too regular to me, so I checked some references -- the 
following may help:

From IEEE-754 (Q754.129 draft, page 49):

8.3 The sign bit

When the sum of two operands with opposite signs (or the 
difference of two operands with like signs) is exactly zero, the 
sign of that sum (or difference) shall be +0 in all rounding 
direction modes except roundTowardNegative; in that mode, the sign 
of an exact zero sum (or difference) shall be –0. However, x+x = 
x–(–x) retains the same sign as x even when x is zero.

--
Cheers,
Kein-Hong Man (esq.)
Kuala Lumpur, Malaysia