Openwall GNU/*/Linux - a small security-enhanced Linux distro for servers
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sun, 13 Sep 2015 22:53:57 +0300
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: Agnieszka's GSoC formats

Hi Agnieszka,

On Sun, Sep 13, 2015 at 09:04:42PM +0200, Agnieszka Bielec wrote:
> I did PR but it's not ideal.
> additionally Lyra2-opencl using new computations of gws and lws
> doesn't work on 960m
> and --test using different passwords doesn't work - BENCHMARK_LENGTH
> 1000 or -1000 or -1001 didn't work
> --test was hunging up when BENCHMARK_LENGTH=1000
> and speed using -1000 or -1001 wasn't the same as using my version for
> making password random, was the same like BENCHMARK_LENGTH=0

Thanks.  Are you going to work on these issues?

I think we don't need to investigate what's going on with those magic
BENCHMARK_LENGTH values, but instead just make all passwords used for
auto-tuning and benchmarking different.  For hashes like this, something
like incrementing a number and using its decimal representation will do.

I think magnum tried to also make the passwords similar to those in the
test vectors.  For some other hash types like md5crypt this makes sense.
So maybe use e.g. increasing decimal numbers padded up to the test
vectors' length?  (And if shorter, do not truncate, as that would result
in duplicate passwords.  Maybe use base64 encoding instead of decimal to
make such occurrences less common.)

Alexander

Powered by blists - more mailing lists

Your e-mail address:

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.