This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
exp(float) much slower than exp(double)
- From: Andreas Svrcek-Seiler <svrci at tbi dot univie dot ac dot at>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Tue, 27 Mar 2007 14:41:08 +0200 (CEST)
- Subject: exp(float) much slower than exp(double)
Hi,
using some simple benchmark I just found out that
exp(float) takes about 8 times more time than exp(double).
gprofing my benchmark gives:
44.26 5.34 5.34 fesetenv
19.49 7.69 2.35 fesetround
14.97 9.49 1.81 feholdexcept
13.40 11.11 1.62 __ieee754_expf
...So most time is wasted(?) in functions never called when the benchmark
is compiled to use doubles.
Sorry if I sent this to the wrong address,
best regards
W. A. Svrcek-Seiler
P.S.:
gcc -v says
Using built-in specs.
Target: x86_64-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man
--infodir=/usr/share/info --enable-shared --enable-threads=posix
--enable-checking=release --with-system-zlib --enable-__cxa_atexit
--disable-libunwind-exceptions --enable-libgcj-multifile
--enable-languages=c,c++,objc,obj-c++,java,fortran,ada
--enable-java-awt=gtk --disable-dssi --enable-plugin
--with-java-home=/usr/lib/jvm/java-1.4.2-gcj-1.4.2.0/jre
--with-cpu=generic --host=x86_64-redhat-linux
Thread model: posix
gcc version 4.1.1 20070105 (Red Hat 4.1.1-51)
(stock fedora core 6, that is)