[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 15 Jun 2012 14:49:25 -0500
From: "jfoug" <jfoug@....net>
To: <john-dev@...ts.openwall.com>
Subject: RE: Re: Re: [patch] optional new raw sha1 implemetation
>5% improvement is nice, but I'm dissapointed it's not higher. Could you
>send me a .o file? I would like to disassemble it and see where GCC goes
>wrong, hopefully there's a simple way to fix it.
This is not a true valid comparison. This is your code, using the gcc
compiler to build the intrinsic sse instructions. The format I compared it
to (raw-sha1), was using a .S file built by ICC. I know the speed from
gcc (cygwin), was 20% or more slower than the icc build (especially for
SHA1). The difference was not as great for MD4/5. But that was identical
code, simply using a different compiler to pre-build the asm.
Jim.
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.
