This is the mail archive of the
glibc-bugs@sourceware.org
mailing list for the glibc project.
[Bug libc/18441] Performance regression due to strcoll_l changes
- From: "carlos at redhat dot com" <sourceware-bugzilla at sourceware dot org>
- To: glibc-bugs at sourceware dot org
- Date: Fri, 22 May 2015 14:43:08 +0000
- Subject: [Bug libc/18441] Performance regression due to strcoll_l changes
- Auto-submitted: auto-generated
- References: <bug-18441-131 at http dot sourceware dot org/bugzilla/>
https://sourceware.org/bugzilla/show_bug.cgi?id=18441
--- Comment #3 from Carlos O'Donell <carlos at redhat dot com> ---
(In reply to Daniel Lichtenberger from comment #2)
> I ran it on Fedora Rawhide (at glibc development release 2.21.90), but it
> still seemed way slower than before.
>
> I will try to come up with a meaningful sample of input data. Is the
> benchmark mentioned in the commit messages available in the glibc source
> tree?
Yes it is. It's the "benchtests" directory under the top-level glibc checkout.
See benchtets/README for more information. The benchmark tests can be run after
a glibc build, so you need to build glibc (but not install it) to run the tests
against your built libraries. In Fedora/RHEL we're working on packaging the
benchmark libraries so you can run them without needing to do a glibc build,
that way you could test your hardware quickly if you found a performance
regression.
To build glibc see:
https://sourceware.org/glibc/wiki/Testing/Builds
Though it should be as easy as configure/make.
--
You are receiving this mail because:
You are on the CC list for the bug.