This is the mail archive of the
glibc-bugs@sourceware.org
mailing list for the glibc project.
[Bug math/13381] rounding problem wih sincosl
- From: "martyn.j.corden at intel dot com" <sourceware-bugzilla at sourceware dot org>
- To: glibc-bugs at sources dot redhat dot com
- Date: Mon, 09 Jan 2012 20:02:37 +0000
- Subject: [Bug math/13381] rounding problem wih sincosl
- Auto-submitted: auto-generated
- References: <bug-13381-131@http.sourceware.org/bugzilla/>
http://sourceware.org/bugzilla/show_bug.cgi?id=13381
martyn.j.corden at intel dot com changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |martyn.j.corden at intel
| |dot com
--- Comment #13 from martyn.j.corden at intel dot com 2012-01-09 20:02:37 UTC ---
Yes, this is likely an issue of slightly different algorithms in the math
library. The main purpose of a sincos function is to give results more
quickly/efficiently that two separate calls to sin and cos, and that might be
better achieved with a different algorithm.
The same issue arose in the math library provided with the Intel
compiler, libimf. Following a customer report, a change was made for version
12.1 of the Intel compiler, (in Intel Composer XE 2011 SP1), so that sincosl
would return the same results as calls to sinl and cosl. Presumably, similar
changes could be made to the GNU libm if this is felt to be an important issue,
otherwise the sincos documentation should be updated to warn of the potential
difference. If you have access to Intelâs libimf, it is also possible to link
to it when compiling with gcc, to work around this issue.
Martyn Corden (Intel Developer Support)
--
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.