This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH v3] powerpc: strstr optimization
- From: Joseph Myers <joseph at codesourcery dot com>
- To: Ondřej Bílka <neleai at seznam dot cz>
- Cc: Carlos O'Donell <carlos at redhat dot com>, Tulio Magno Quites Machado Filho <tuliom at linux dot vnet dot ibm dot com>, GNU C Library <libc-alpha at sourceware dot org>, Steve Munroe <sjmunroe at us dot ibm dot com>, Rajalakshmi Srinivasaraghavan <raji at linux dot vnet dot ibm dot com>
- Date: Mon, 27 Jul 2015 14:51:35 +0000
- Subject: Re: [PATCH v3] powerpc: strstr optimization
- Authentication-results: sourceware.org; auth=none
- References: <558A5642 dot 5020107 at linux dot vnet dot ibm dot com> <558A5761 dot 2000409 at linux dot vnet dot ibm dot com> <87oajpm8nc dot fsf at totoro dot br dot ibm dot com> <871tgijuri dot fsf at linux dot vnet dot ibm dot com> <55A6FE3F dot 6090701 at redhat dot com> <55A70B70 dot 6090607 at redhat dot com> <20150716195538 dot GA5140 at domone> <55A8110C dot 7000209 at redhat dot com> <alpine dot DEB dot 2 dot 10 dot 1507221607370 dot 21570 at digraph dot polyomino dot org dot uk> <20150725082025 dot GA5043 at domone>
On Sat, 25 Jul 2015, OndÅej BÃlka wrote:
> Joseph, what is policy about assembly implementations when there is
> better generic implementation? This assembly is slower than when do
Removal of such an assembly implementation is a simple matter of getting
consensus, typically based on benchmark results from the checked-in
benchtests (so if necessary, you may need to get consensus on an addition
to such benchtests first). The most likely form of such consensus is
through agreement from the architecture maintainer, but I don't expect
such changes, properly supported by benchmark results from the checked-in
benchtests, to be controversial.
Of course the comparison is with the *checked-in* generic implementation,
not something posted but not checked in.
> simple optimization of C strstr, with patch. I repeatedly asked to
> check it but never got reply.
>
> [PATCH] Improve generic strstr performance.
Well, if a patch takes a while to get reviewed, keep pinging, while making
sure that the patch submission includes all necessary information, that
the required benchmarks have been reviewed and checked in first, etc.
You have about 80 unreviewed patches showing in the patchwork list. If
any of those are superseded by later versions of the same patch, I
strongly advise marking the earlier versions as superseded in patchwork so
that anyone using patchwork to find patches to review finds those patches
of yours that actually need review and not a load of noise from older
superseded patch versions. (The same applies to anyone submitting patches
- keep the list of your patches pending review clean - but your list is
particularly long.)
--
Joseph S. Myers
joseph@codesourcery.com