Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 09 May 2011 08:28:30 +0200
From: magnum <>
Subject: Re: John core change patch (and md5-gen, etc)

On 2011-05-09 02:50, JimF wrote:
> I have a pen drive with Unbutu on it now, so can more easily test
> intrinsic code, and get it fully working. But this may take some
> coordination again, to get a good stable version which builds properly
> if a user WANTS intrinsic code, but also builds properly if they do not
> (such as using GCC, or 32 bit systems).

You *want* intrinsics even if using gcc, on 64-bit, if you meant 
otherwise. Last I checked (gcc 4.5.1 vs icc and 
1.7.6-Jumbo-12+intrinsics+fixes) I got this:

gcc w/o intrinsics: 10832 c/s
clang w/intrinsics: 18108 c/s
gcc w/ intrinsics:  19284 c/s
icc w/ intrinsics:  28608 c/s

gcc w/o intrinsics:  2321 c/s
clang w/intrinsics:  9744 c/s
gcc w/ intrinsics:  10296 c/s
icc w/ intrinsics:  16152 c/s

gcc w/o intrinsics:  4640K c/s
clang w/intrinsics: 15711K c/s
gcc w/ intrinsics:  16167K c/s
icc w/ intrinsics:  23509K c/s

Here is an example where only icc performed much better. If I understand 
things right, This format do not have intrinsics but icc managed to do 
fine without them, right?

gcc w/o intrinsics: 13965K c/s
clang w/intrinsics: 13918K c/s
gcc w/ intrinsics:  13937K c/s
icc w/ intrinsics:  30502K c/s


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.