This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH][PPC] Remove unnecessary local variable mptwo
- From: "Ryan S. Arnold" <ryan dot arnold at gmail dot com>
- To: "Carlos O'Donell" <carlos at systemhalted dot org>
- Cc: Siddhesh Poyarekar <siddhesh at redhat dot com>, libc-alpha <libc-alpha at sourceware dot org>, "Ryan S. Arnold" <rsa at us dot ibm dot com>
- Date: Thu, 10 Jan 2013 10:50:27 -0600
- Subject: Re: [PATCH][PPC] Remove unnecessary local variable mptwo
- References: <20130109114033.GC2692@spoyarek.pnq.redhat.com><50EEE333.7080304@systemhalted.org>
On Thu, Jan 10, 2013 at 9:50 AM, Carlos O'Donell
<carlos@systemhalted.org> wrote:
> On 01/09/2013 06:40 AM, Siddhesh Poyarekar wrote:
>> Hi,
>>
>> Attached patch removes the local variable mptwo from the __inv
>> function so that it uses the global static definition instead. No
>> regressions resulting from this patch. OK to commit?
>>
>> Siddhesh
>>
>> * sysdeps/powerpc/powerpc32/power4/fpu/mpa.c (__inv): Remove
>> local variable MPTWO.
>> * sysdeps/powerpc/powerpc64/power4/fpu/mpa.c (__inv): Likewise.
>
> How does this impact the generated code?
>
> How does that impact mpatan's performance?
Preventing regressions in performance of mpa is certainly something we
(IBM) are interested in. In-fact we'd like to see improvements in
performance of this code if possible.
I understand that these cleanups are ultimately 'a good thing' and we
do have time to measure performance of the 2.17 release before we
release 2.18.
I think in order to have a good measure of the performance impacts we
need to make sure that Siddhesh's performance benchmark suite is in
place and that it can work with the 2.17 release so that we can get a
baseline.
Ryan