[Bug math/28472] pow(10, i) accuracy

["wdijkstr at arm dot com" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=28472

--- Comment #22 from Wilco <wdijkstr at arm dot com> ---
(In reply to b. from comment #21)
> dear guys,  
>   
> in a way you are talking this point to death,  
>   
> this issue is not about powers of two, usually they have   
> few problems in binary datatypes, but about powers of 10!  
>   
> it's important to have them exact ( as good as possible )  
> at integral powers of ten,  

We do get all of the powers of 10 that are exactly representable correct. And
almost all of the inexact cases are rounded correctly.

> and regarding monotonity it's surely better to smoothe the  
> nearby values to the exact values than the other way.  
>   
> Note, we need 2! results changed for binary64's ( bin80's  
> are worse ) and there are no monotonity issues around them.

There isn't an obvious fix for the current implementation. If you can find a
way without a performance hit, I'd love to hear it. If you want correctly
rounded results (and don't care about performance), just use a correctly
rounded math library.

-- 
You are receiving this mail because:
You are on the CC list for the bug.