Mathscipt node issue

[dungnnguyen]

I did. They submitted to R&D and closed the case.

[altenbach]

---

@dungnnguyen wrote:

I did. They submitted to R&D and closed the case.

---

OK, there must be a CAR# in their system to track the bug. I guess you never got the number...

[altenbach]

OK, I wrote a pure g replacement and it seems more robust.

Here is some simple code to calculate N and D for any input X such that N/D ~= X. See if this works for you.

(I also included a little tester to measure performance and failure count. Note that this is just a draft. Once polished up, I will probably post it in a dedicated place elsewhere. For typical tolerances of 1e-6, the computation takes about 500ns).

[BoKnows]

CAR 702329

[dungnnguyen]

Thanks, it's matched matlab in all the cases I tested.