cubic root (cbrt) fails
Issue #152
closed
Originally reported on Google Code with ID 152
What steps will reproduce the problem?
1. cbrt(1.09)
What is the expected output? What do you see instead?
expected: 1,0291424665715064607373575413984
seen: 1
What version of the product are you using? On what operating system?
v0.9, windows xp sp2
Please provide any additional information below.
cbrt(1.1) is ok
cbrt(1.099999999999999999) fails
cbrt(1.000000000000000001) fails
cbrt(1) is ok
Reported by gabriel.gross
on 2008-02-03 22:14:53
Comments (4)
-
Account Deleted -
Account Deleted Reported by `wolf.lammen` on 2008-02-04 02:32:25
-
repo owner - changed component to mathengine
- edited description
- changed milestone to 0.10
- changed version to 0.9
-
repo owner - changed status to closed
- Log in to comment
``` confirmed and fixed in 0.9 and trunk. The Newtorn-Raphson loop started with a fixed working precision. But for values 1+x with x << 1 the working precision in the first iteration has to be adjusted such that the small changes produced by the small x are not ignored. ```
Reported by `wolf.lammen` on 2008-02-04 02:28:46 - Status changed: `Fixed` - Labels added: Milestone-0.10