[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20150217211247.GI23507@brightrain.aerifal.cx>
Date: Tue, 17 Feb 2015 16:12:47 -0500
From: Rich Felker <dalias@...c.org>
To: Denys Vlasenko <vda.linux@...glemail.com>
Cc: musl <musl@...ts.openwall.com>
Subject: Re: [PATCH] x86_64/memset: use "small block" code for blocks
up to 30 bytes long
On Tue, Feb 17, 2015 at 07:53:28PM +0100, Denys Vlasenko wrote:
> On Tue, Feb 17, 2015 at 6:40 PM, Rich Felker <dalias@...c.org> wrote:
> >> With your program I see similar results:
> >>
> >> ....
> >> size 50: min=10, avg=10 min=10, avg=10
> >> size 52: min=10, avg=10 min=10, avg=10
> >
> > The ... was the part where mine seemed better. :)
>
> Do you seriously think I would go as low as lying by omission?
> Here are the full, unabridged files of three runs of both algorithms.
No, that's not what I meant, and I'm sorry for making it sound that
way. I just meant it was the interesting part I wanted to compare. For
me, I get ~1 cycle difference for a number of the small runs, but on
your test that measures rates in a way that's probably more applicable
to real-world use, there was no measureable difference one way or the
other.
Rich
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
