phc-discussions - Re: [PHC] OMG we have benchmarks

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Wed, 1 Apr 2015 05:22:43 -0500 (CDT)
From: Steve Thomas <steve@...tu.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] OMG we have benchmarks

> On April 1, 2015 at 4:18 AM Gregory Maxwell <gmaxwell@...il.com> wrote:
>
> On Wed, Apr 1, 2015 at 9:14 AM, Krisztián Pintér <pinterkr@...il.com> wrote:
> > On Wed, Apr 1, 2015 at 11:01 AM, Milan Broz <gmazyland@...il.com> wrote:
> >> - the low memory setting is "unstable" because of RUSAGE measurement:
> >> Real memory us is simple difference of getrusage(RUSAGE_SELF, ...)
> >> before and after run (well, here maximum of three runs).
> >
> > just a quick question: wouldn't it be easier and more precise to
> > calculate the memory requirement instead of measuring it? it should be
> > quite straightforward from the algorithms.
>
> And would also avoid making mistakes like crediting an algorithm for
> more than it actually needs due to malloc overhead or the like. It's
> not bad to look at the rusage data since it should always be higher
> than the actual usage, and might catch some mistake... but it
> shouldn't be the definitive figure as an attacker (or mature
> implementation) is going to happily optimize out whatever excess
> overheads you're accidentally counting with it.

For the most part with at least 8 MiB they line up. Catena-Butterfly is the odd
ball.