Hey,
Apologies if this question is daft; I'm new to both hashcat and the field in general.
I understand that the work carried out by hashcat is parallel (or at least the generation of candidate keys is) and that GPUs are particularly well suited to the task.
Given the prices of GPUs at the moment and the need for parallel working, has anyone tried spreading the work over several Raspberry Pi machines?
I was messing around with my own take on password cracking in a .net project (it's slow - it was just a PoC) and I figured out pretty quickly that it's way quicker to set a thread-per-core for a given set of candidates.
i.e. if you had ?l?l?l?l (a password known to be four lower case characters) and 8 cores then the first set of threads would be
a?l?l?l
b?l?l?l
c?l?l?l
d?l?l?l
e?l?l?l
f?l?l?l
g?l?l?l
h?l?l?l
and given the price of a Pi at the moment one could have 16 cores + 8gb Ram (Pi 4 B with 2GB Ram) for ~$200.
I had a look at hashcat brain, but I don't know that this would do what we want. I'd imagine you'd want one "server" dishing out the candidates and all the other Pis just consuming the candidates until one of them gets lucky (or, if --keep-guessing, all candidates are used up)
Apologies if this question is daft; I'm new to both hashcat and the field in general.
I understand that the work carried out by hashcat is parallel (or at least the generation of candidate keys is) and that GPUs are particularly well suited to the task.
Given the prices of GPUs at the moment and the need for parallel working, has anyone tried spreading the work over several Raspberry Pi machines?
I was messing around with my own take on password cracking in a .net project (it's slow - it was just a PoC) and I figured out pretty quickly that it's way quicker to set a thread-per-core for a given set of candidates.
i.e. if you had ?l?l?l?l (a password known to be four lower case characters) and 8 cores then the first set of threads would be
a?l?l?l
b?l?l?l
c?l?l?l
d?l?l?l
e?l?l?l
f?l?l?l
g?l?l?l
h?l?l?l
and given the price of a Pi at the moment one could have 16 cores + 8gb Ram (Pi 4 B with 2GB Ram) for ~$200.
I had a look at hashcat brain, but I don't know that this would do what we want. I'd imagine you'd want one "server" dishing out the candidates and all the other Pis just consuming the candidates until one of them gets lucky (or, if --keep-guessing, all candidates are used up)