Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Tue, 24 Jan 2012 00:12:01 +0100
From: magnum <john.magnum@...hmail.com>
To: john-dev@...ts.openwall.com
Subject: Re: gcc versions

On 01/05/2012 12:56 PM, Solar Designer wrote:
>> BTW, would you 
>> be able to try my testpara make targets with all those versions of gcc?
> 
> I just did - again, for x86-64 only.  Here are the results.
> Surprisingly, gcc 4.0.0 performed just fine - no need to disable these
> intrinsics for it.  The 4.4.0 vs. 4.4.6 difference is curious.  I may
> need to make more 4.4.x builds.

We have now had the SHA-1 reworked by JimF. I believe the "para" values
you collected will work fine but ideally you should run that again
if/when possible (I presume it was scripted :)

The latest testpara also compares old 80x4 SHA-1 intrinsics and new 16x4
one (but so far the new code is faster in all situations). For 32-bit
builds, it also shows the asm versions' speeds.

magnum

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.