[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20150916194044.GA12655@openwall.com>
Date: Wed, 16 Sep 2015 22:40:45 +0300
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: fast hash processing bottlenecks (was: ldr_split_line() performance regression)
On Wed, Sep 16, 2015 at 10:31:35PM +0300, Solar Designer wrote:
> The potential speed for the current raw-md5 code on this machine is:
>
> Benchmarking: Raw-MD5 [MD5 128/128 SSE4.1 4x3]... DONE
> Raw: 21811K c/s real, 21811K c/s virtual
>
> per core. But we don't reach anywhere near it with wordlist mode.
> We do reach 21M+ per core with --fork=8 e.g. in mask mode. 20M for
> incremental mode locked to length 8. Also, 19.7M for wordlist+mask
> (tested with -mask='?w?a?a'). But not for wordlist+rules.
I've just tried -mask='?w?a' instead of -ru=best64 on the 29M hashes.
It ran for about the same 47 seconds, but tested ~1.5x more passwords
(equivalent to 95 rules, rather than 64). The per core c/s rate was
around 7M, up from around 4.5M for rules (also when running the 29M).
I think ~1.5x for this testcase is roughly the speedup we can expect
from implementing a "rules first" mode. (And the speedup would be
greater when running against few hashes.)
Alexander
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.
