Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 16 Sep 2015 04:28:42 +0300
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: Judy array

On Wed, Sep 16, 2015 at 02:29:17AM +0300, Solar Designer wrote:
> real    0m57.785s
> user    4m21.243s
> sys     0m15.629s
> 
> Oops.  Almost same speed without actual prefetching.  So maybe the
> re-arranged code is somehow faster in this case, even when I don't
> prefetch.  Or maybe the CPU does speculative execution and thus
> prefetches anyway, using those pointers that the following few loop
> iterations would use (especially as the compiler unrolls the loop).

Implemented better prefetching, patch attached.  New speed:

real    0m56.260s
user    4m9.024s
sys     0m16.240s

Best of several runs:

real    0m55.987s
user    4m7.024s
sys     0m16.082s

I think this should be committed, and I might get it into core tree
later.  Even though we'll probably replace this with Fred's code soon.
I guess what I am doing is providing another baseline for comparison.

Alexander

View attachment "john-huge-prefetch.diff" of type "text/plain" (4518 bytes)

Powered by blists - more mailing lists

Your e-mail address:

Powered by Openwall GNU/*/Linux - Powered by OpenVZ