This is the mail archive of the
glibc-bugs@sourceware.org
mailing list for the glibc project.
[Bug math/706] pow() produces inaccurate results for base ~ 1.0, and large exponent on 32-bit x86
- From: "vincent-srcware at vinc17 dot net" <sourceware-bugzilla at sourceware dot org>
- To: glibc-bugs at sources dot redhat dot com
- Date: Tue, 28 Feb 2012 14:03:00 +0000
- Subject: [Bug math/706] pow() produces inaccurate results for base ~ 1.0, and large exponent on 32-bit x86
- Auto-submitted: auto-generated
- References: <bug-706-131@http.sourceware.org/bugzilla/>
http://sourceware.org/bugzilla/show_bug.cgi?id=706
--- Comment #12 from Vincent LefÃvre <vincent-srcware at vinc17 dot net> 2012-02-28 14:03:00 UTC ---
(In reply to comment #11)
> What I proposed is just an easy workaround which can be easily and quickly
> implemented (and it would improve the actual results).
I'm not sure that improving the code for integer exponents only would be a good
idea. A consequence would be that the f(x) = pow(a,x) implementation would be
completely non-monotonous (with important differences near the integer inputs),
with possible bad side effects on some codes.
> About the char in the C code: it is just used for its sign (+/- 1) not for its
> value.
But a sign is part of the value. On platforms where char is unsigned, a char
value cannot have a negative sign.
--
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.