phc-discussions - Re: [PHC] Best RNG for filling memory?

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]

Message-ID: <52BB31B1.7010003@dei.uc.pt>
Date: Wed, 25 Dec 2013 19:27:45 +0000
From: Samuel Neves <sneves@....uc.pt>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] Best RNG for filling memory?

On 25-12-2013 03:51, Bill Cox wrote:
> Sorry for this poorly informed post... I was unable to find a good
> source package for benchmarking salsa20.  Do you know where I can get
> one that simply spits out random numbers for a stream cipher that I
> can easily benchmark?

Probably the best resource, as pointed out by Tony, is eBACS:
http://bench.cr.yp.to/results-stream.html. The times are measured in
cycles per byte, so lower is better.

hydra7 (Intel Sandy Bridge) seems to have AES-NI implementations, which
are still slower than Chacha8:
http://bench.cr.yp.to/results-stream.html#amd64-hydra7

>
> MARC speeds up ARC4 by almost 2X, making it speed competitive with
> Blake2.  Blake2 is based on ChaCha, which in turn is based on Salsa20.
>  From what I've read so far, ChaCha claimed to be faster than Slasa20,
> and Blake claimed to be based on ChaCha.  Why would Salsa20 be faster
> than Blake2?
BLAKE2 is a hash function. It has different security requirements, which
end up requiring more computation (i.e., it's slower) than a dedicated
stream cipher like Salsa20. How are you even using it as a PRNG? It can
be used to make one, but that's not the first option I'd consider.