VI of the Day (8/24/2009)

[Ben]

Well I blew the chance at teh bonus points now that I have more clues. (Oh bother!)

I have another idea!

We have done everything but actually say the word "epsilon" but now I have to.

I found this code*

 

inthis thread that was originally posted by Charles K of NI.

Are we getting close?

Ben

* This is slight mode of teh original code to toss a ms wait and eleminate the candidate display update in each iteration.

Message Edited by Ben on 08-27-2009 08:17 AM

[DFGray]

Getting warmer.  Hint number 2 - what are the outputs of the VI?

[Ben]

I'm tossing in the towel!

It must have something to do with the "measured Precision" output but I don't know what that value is supposed to mean.

Ben

[F._Schubert]

+1 must come after the (-x)

Felix 

[Ben]

---

F. Schubert wrote:

+1 must come after the (-x)

Felix 

---

I was never very good at memorizing numbers so I'll trust you on that point.

Still playing with that example and was thinking about a "effectivly equal" VI that takes epsilon into account. So I changed the control to a contnat thinking that LV would fold the entire VI... but NO!

 

Why wouldn't LV fold that loop?

Ben

[dan_u]

---

Ben wrote:

Why wouldn't LV fold that loop?

---

If I remember correctly I've never seen LV folding a while loop!?

The question remains though: why?

Daniel

Message Edited by dan_u on 08-27-2009 04:14 PM

[Ben]

---

dan_u wrote:

---

Ben wrote:

Why wouldn't LV fold that loop?

---

If I remember correctly I've never seen LV folding a while loop!?

The question remains though: why?

Daniel

Message Edited by dan_u on 08-27-2009 04:14 PM

---

While review the 2009 complete upgrade notes I was looking for how to use the new icon editor when the notes about floding caught my eye.

I had debugging enabled on that VI so no folding. Disable debugging and the while loop is folded.

But what to my wondering eyes did not appear, the VI snippets leave out folding I fear!

here is a screen shot.

 

Ben

[DFGray]

For the record, the posted snippet is not <vi.lib>\Waveform\WDTOps.llb\Check for Equality.vi.  That would be this (for some strange reason, this VI is diagram locked on earlier version of LabVIEW).

The glaring error is shown on the inner frame.  The measured precision output, which is to how many digits the input numbers match, should be 15 (it is almost 16, log₁₀(2⁵³)) for two equal doubles.  It is set to 10.