Looking for TVM contributions

[robve]

(06-15-2024 03:39 AM)Albert Chan Wrote:  BTW, I think your Newton's code termination test is wrong.
It does not look like HP71B LINE 80 at all!

Your wish is my command! I changed the termination test to yours. The results are much worse:


  tvmperfcomp.pdf (Size: 36.85 KB / Downloads: 7)

EDIT: I should mention that the Newton results in the PDF above use i+h*h>i as was used in Albert's HP-71B code before. Actually, I also tried using fsign(y) == fsign(w) && fabs(y) < fabs(w) as shown n Albert's post above but then Newton only iterates once so the results are much worse. I don't understand why that works on the HP-71B. I'm probably doing something stupid.

- Rob