hashcat Forum

Hello everyone,

I have a spare Gpu rig where I have correctly configured Hashcat / Ubuntu 18.04 Server and it is working fine.

This rig have a really cheap processor, therefore there is not enough power to produce enough work to load all the Gpus at 100%.

Few month ago I have build a Raspberry Pi B3 + Cluster and my idea is to use it to generate the wordlist and pipe it straight to the rig, doing so hopefully I will be able to remove the processor bottleneck and let the Gpus run much faster.
 
Actually I am trying to write different programs in c to build up some wordlist, but being honest try to use all the resources of the cluster is not easy at all , and maybe someone already did it better than how I could.

At this point I would like to understand if there is any feature of Hashcat or any program around that is going to be suitable for my intent.

Have anyone used a Cluster before for do something like that? 
Maybe someone can give me some advise...


My rig

12 x Amd ( rx 470,480,570,580 4/8 Gb)
Intel G 4400 3.30 Ghz
TB250-BTC PRO
16 Gb DDR4
Evga G3 1000w
2 x Hp 1200w Server


Thanks a lot to everyone and have a lovely day

you're assuming your ethernet network connection is faster than your cpu/cpu-bus?

(07-05-2019, 07:37 PM)undeath Wrote: [ -> ]you're assuming your ethernet network connection is faster than your cpu/cpu-bus?

To be honest I don't even have thinker about that, And it is a really good point , and a mistake on my side.

The point is that if I try to generate the wordlist with the processor and pipe it in to  hashcat it cannot load the gpus at 100%, but if I use an already made wordlist from the ssd the Gpus are running at 100 % constantly ( in another rig that I have with 6 gpus and a i7 4770 k).

For this reason The idea was build the wordlist with the Cluster and then use the processor just for pipe them to the Gpus.

I just saw the Cluster in the corner of my room and I want to try to use it ,  I didn't find any documentation about it, only people that were computing the hashes in the Cluster.

So I can imagine that the only solution is a double socket motherboard.

Otherwise ,is possible to use one of the Gpus to generate the wordlist for the other gpu?

**Edited

(07-05-2019, 07:29 PM)Ryukxx88 Wrote: [ -> ]Few month ago I have build a Raspberry Pi B3 + Cluster

What do you mean by Cluster? What hardware is the cluster made of? :)