This is the mail archive of the libc-hacker@sources.redhat.com mailing list for the glibc project.

Note that libc-hacker is a closed list. You may look at the archives of this list, but subscription and posting are not open.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

inlined string tests are slower on i686


This is hardly a complete benchmark, but here goes:

inlined:

$ /usr/bin/time string/inl-tester >/dev/null
0.77user 0.00system 0:00.77elapsed 100%CPU

$ /usr/bin/time string/inl-tester >/dev/null
0.77user 0.00system 0:00.77elapsed 100%CPU

$ /usr/bin/time string/inl-tester >/dev/null
0.77user 0.00system 0:00.77elapsed 100%CPU

non-inlined:

$ /usr/bin/time string/noinl-tester >/dev/null
0.74user 0.00system 0:00.74elapsed 99%CPU

$ /usr/bin/time string/noinl-tester >/dev/null
0.74user 0.00system 0:00.74elapsed 99%CPU

$ /usr/bin/time string/noinl-tester >/dev/null
0.74user 0.01system 0:00.74elapsed 101%CPU

The non-inlined version is 4% faster

Does anyone have more comprehensive benchmarks that justify string
inlines for ix86?

Does anyone know how i[345]86 perform in this regard?

Greg

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]