Complex matrix inversion with issues
Re: Complex matrix inversion with issues
Thanks Thomas. I was already very satified with this 68. But your are right ... it's their software ... I suppose they know what they are doing.
-
- Posts: 1107
- Joined: Tue May 02, 2017 5:48 pm
- Location: Netherlands
- Contact:
Re: Complex matrix inversion with issues
I explained my reasoning for using 68 to Mr. Cornea; he repeated it should be 69 but didn't explain why. Oh well. I made the change and it will be in the next official release (Free42 3.0.1 has my work-around which avoids bid128_hypot() altogether). The patch using 68 isn't bad, of course, if Marius Cornea is right about 69 then my version can potentially lose 1 digit of precision, which is still a lot better than the original which can lose 33. (Or is it 32? Or 34? Aaaah )