Solar Designer

@sectroyer We have the `--node` option, usable with any cracking mode (although under the hood it's implemented with separate code per cracking mode). If your nodes aren't the same speed,...

Anyhow, there isn't anything in what you wrote that we'd currently want to track as an issue. So I'll close this. Please feel free to add further comments anyway, or...

> Unless their keyspace numbers are NOT compatible with how you currently handle splitting by node parameter? They almost certainly are not. BTW, you mention "brute force" in multiple places,...

@sectroyer The disappearing progress percentage and ETA could be a bug or at least a shortcoming. If you can share a reduced and reproducible test case for this, then please...

@sectroyer Thank you for creating that other issue. There isn't currently a good way to express those masks as one attack. BTW, for attacks on a small number of descrypt...

@sectroyer Now we're deep into topics that would be much better discussed on john-users than here. > I am MUCH MORE interested in adjusting the system itself and evenly distributing...

> At least those are results I have after analysing some IoT hashes I played around. It's a fallacy to focus on password patterns you see the most without considering...

@sectroyer More john-users material here. Please use the mailing list next time you bring something like this up. > best64 rule (btw is there any other rule also worth checking?)...

A further and more important optimization for 2i would be to introduce index mapping and store only those blocks that would actually be subsequently used, reusing memory for blocks that...

Also relevant: Towards Practical Attacks on Argon2i and Balloon Hashing [PDF](https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7961977) [Joël Alwen](https://ieeexplore.ieee.org/author/37086121535); [Jeremiah Blocki](https://ieeexplore.ieee.org/author/37947782900) https://ieeexplore.ieee.org/document/7961977 but focuses on larger m_cost than likely practical for GPUs anyway (1+ GB).