With bitcoin/litecoin having shifted to ASIC's a large number of GPU mining rigs have been pressed into service as crypto clusters for oclHashcat. Unfortunately what works for bitcoin/litecoin is terrible as a hashcat farm. This talk goes over the history of a large cluster, infrastructure required to keep it running, hardware build standards, and lessons learned from large amounts of failure to build a highly successful cluster. This talk will also discuss ROI and is basically the anti werby talk (more cost effective to use dedicated rigs rather than the cloud) This talk will also discuss the sourcing of various parts of the cluster and detail current clustering software in use inclduing how to spec a cluster control node.

Member of Team Hashcat, Security Assurance Manager (Red Team) for a government owned financial institution. Spent most of career in aerospace/defense, has run a large GPU cluster for over 5 years, current cluster size is ~60 GPU's across a dozen nodes. Likes beer, toast, and plastic army men with parachutes.

@EvilMog

Back to Derbycon 2015 video list