This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] powerpc64: strrchr optimization for power8
- From: Rajalakshmi Srinivasaraghavan <raji at linux dot vnet dot ibm dot com>
- To: "Carlos O'Donell" <carlos at redhat dot com>, libc-alpha at sourceware dot org
- Date: Tue, 21 Mar 2017 10:44:03 +0530
- Subject: Re: [PATCH] powerpc64: strrchr optimization for power8
- Authentication-results: sourceware.org; auth=none
- References: <1487070321-27700-1-git-send-email-raji@linux.vnet.ibm.com> <52479e2b-f675-9d6e-7b34-27e24ee48081@redhat.com> <55dfbdb0-3776-caa9-e87b-10c297b4b496@linux.vnet.ibm.com> <139ea9b7-005d-fd33-a223-fe36fc995f4d@redhat.com> <00958c35-0b2d-60c9-f618-cfcd185710e8@linux.vnet.ibm.com> <5af3baaf-a340-cf07-10fc-f877d803dd6c@linux.vnet.ibm.com> <3e1454d4-ce03-f3a5-ac6f-6d9a7a0151f8@redhat.com>
On 03/17/2017 09:08 PM, Carlos O'Donell wrote:
Knowing the mean value of string lengths would let us decide where
to place most of optimization efforts. I don't know that we have any
good references to academic literature here.
No, I dont have details about strrchr call statistics. This
optimization is to make use of POWER 8 capabilties in general.
Now I plotted the power8 performance and there is a big bump in the middle,
any idea why?
I could see some sudden increase for sizes 104-118 and around 222
in the shared sheet. However there is no special logic in the code
related to these sizes. I tried to check if this happens on my
test P8 ppc64le system and I could not recreate this.
--
Thanks
Rajalakshmi S