Rounding

[Cloedu]

Hello,

i have found unexpexted rounding effects with the function "Round to nearest integer"

Input 0.5

output 0

Input 1.5

output 2

Input 2.5

Output 2

Input 3.5

Output 3

Labview is alwas rounding to a even integer.... but why?

What is the sense of these implementation of auf Rounding alghorithm?

Thanx

Claude

[LAVA 1.0 Content]

This is standard behavior and noted in the help. I've just written a wiki article on rounding.

Ton

[tnt]

QUOTE (Cloedu @ Aug 25 2008, 12:00 PM)

Labview is alwas rounding to a even integer.... but why?

Hi,

as far as I remember this had to do with statistics,

assume you have a big pile of points and they all get rounded up when ending on .5,

you end up with a higher average,

with this statistical rounding, the average would not change that much.

I hope this helps a bit

edit: Ton was a bit quicker in typing as me...

[jgcode]

QUOTE (Cloedu @ Aug 25 2008, 06:00 PM)

i have found unexpexted rounding effects with the function "Round to nearest integer"

Thats how I was taught in school.

[mross]

QUOTE (Cloedu @ Aug 25 2008, 06:00 AM)

Hello,

Labview is alwas rounding to a even integer.... but why?

What is the sense of these implementation of auf Rounding alghorithm?

Thanx

Claude

LabVIEW follows IEEE standard

round up the odd whole numbers ending with .5,

round down even numbers ending with .5.

From message by

Lew Drake

Process Automation Corporation, and the info LabVIEW mailing list, 4/12/07.

This method accumulates less error over time, it is statistically defensible. If you look around on line you will find detailed explanations, no problem.

[Yair]

If you want to implement common rounding, you can do this by adding 0.5 and then rounding down. Here's an example.