Posts: 455
Threads: 39
Joined: Jan 2011
I have written a little program to "benchmark" a few of my calculators. It simply adds all the integers from the start value down to 1. I am using x=0? to determin when to stop.
Out of interest I wanted to see what happens if I put the calculator into integer mode and discovered something that I do not understand.
x=0? returns "true" for x being -75. Can someone explain why that is?
According to the manual I should use x=+0? or x=-0?, but haven't tried that yet (forgot the calculator at home)
Cheers,
Harald
Edited: 13 Mar 2012, 9:48 a.m.
Posts: 228
Threads: 7
Joined: Aug 2011
notwithstanding what the manual says about x=-0 and x=+0, x=0? should still work as expected in integer mode.
Not only does -75 x=0? return true, 0 x=0? returns false!
Posts: 4,587
Threads: 105
Joined: Jul 2005
Good find :-) Strange things happening there: -16 x=0? returns (correctly) false, so what's happening in between? We'll look into that. Thanks for reporting!
Posts: 3,283
Threads: 104
Joined: Jul 2005
I have an idea where to look but I need time to investigate it.
Posts: 3,229
Threads: 42
Joined: Jul 2006
Fix committed but not built. This was a result of the rearrangement of the internal constants table recently. There was a second bug where the integer test against 1 was really against -1 :-(
You only need to use x=+0? and x=-0? in special occasions and if you don't know or understand when, you don't need to use them :-) Just use x=0? instead.
- Pauli
Posts: 455
Threads: 39
Joined: Jan 2011
Quote:
You only need to use x=+0? and x=-0? in special occasions and if you don't know or understand when, you don't need to use them :-) Just use x=0? instead.
- Pauli
So basically you are saying they are only interesting if I am using ones' complement because then I have both +0 and -0? But x=0? should be true for both +0 and -0?
Posts: 3,229
Threads: 42
Joined: Jul 2006
Quote:
So basically you are saying they are only interesting if I am using ones' complement because then I have both +0 and -0? But x=0? should be true for both +0 and -0?
That is a third of the ways negative zero can appear.
The sign & mantissa integer mode also has positive and negative zero. This isn't really any different to the 1s complement mode however.
Real numbers have positive and negative zeros, which is invaluable for numerical programming. You'll have to set flag D to see the difference.
And yes, x=0? is true for both positive and negative zero.
- Pauli
Posts: 455
Threads: 39
Joined: Jan 2011
Quote:
Real numbers have positive and negative zeros, which is invaluable for numerical programming. You'll have to set flag D to see the difference.
I never thought about that, but I suppose it helps to know which side you are approaching a limit from.
Quote:
And yes, x=0? is true for both positive and negative zero.
- Pauli
Ok, thanks for the info!
Harald
Posts: 3,229
Threads: 42
Joined: Jul 2006
Quote:
...I suppose it helps to know which side you are approaching a limit from.
Exactly. Very useful for complex branch cuts e.g.
- Pauli