The Cryptohaze tools are a set of GPU accelerated password cracking tools. There are two main tool sets. The Cryptohaze Multiforcer is a network-enabled password brute forcing tool using CUDA for nVidia GPUs. The Cryptohaze Rainbow Table tools are a fully GPU accelerated set of rainbow table tools for both CUDA and OpenCL. They are tested on nVidia GPUs, AMD/ATI GPUs, and CPUs (using Intel OpenCL SDK on Intel, and the AMD OpenCL SDK on AMD).
All the releases work on Windows, Linux, and OS X. The binary releases are 64-bit only, although the tools can be built 32-bit. The Multiforcer is fine in 32-bit code, but the rainbow table utilities absolutely will not work on anything but toy problem sizes with 32-bit code, as they extensively use memory mapping to load files, and 32-bit code cannot map in more than roughly 2GB of data at once.
Downloading the tools
You can download binary 64-bit releases for your platform here: http://sourceforge.net/projects/cryptohaze/files/Cryptohaze-Combined/
Running the tools on EC2
If you don't have convenient GPUs at home, you can use EC2! Cryptohaze on EC2
The Cryptohaze Multiforcer is an easy to use password brute forcing utility. It supports network distribution of work so multiple systems can work on the same problem, joining and departing as needed without losing work. The tool supports a wide variety of hashes, and supports multiple GPUs in a system. Note: The Cryptohaze Multiforcer is CUDA only! See the MFN framework for the OpenCL version.
Cryptohaze Multiforcer : More details and how to use the tool.
Cryptohaze Multiforcer Supported Hashes : A list of the currently supported hash types.
MFN (Multiforcer New)
The MFN framework is the new revision of the Multiforcer. This contains improvements developed in the last 5 years of work, and is much more modular. In addition, it supports both CPU (SSE2) and OpenCL (for AMD GPUs) support! This is a ground up rewrite designed to significantly improve performance, and to make it easier than ever for developers to add new hash algorithms & capabilities. Yes, developers! It's pretty much the only remaining open source GPU-focused password cracking toolchain out there (yes, there's John, but it's still fundamentally a CPU-focused tool).
The MFN tool is in beta now, and will hopefully be hitting RC status soon!
Cryptohaze Rainbow Tables
Cryptohaze Rainbow Tables are a fully GPU accelerated implementation of the Rainbow Tables concept. The tools are cross platform and work with whatever you have - nVidia GPUS with CUDA, ATI GPUs with OpenCL, and both Intel & AMD CPUs with OpenCL. They are well suited to cracking small numbers of passwords quickly, and custom tables can be generated if you have specific needs.
Cryptohaze GPU Rainbow Tables : More detailed overview and tool details.
Cryptohaze WebTables: A way to search tables without having to download them first! Information on how to use this and free servers.
If you are interested in contributing to the project, consider looking at the Developer Guide
One thing that may be very useful (that is a work in progress) is the Developer Concepts Guide - this is an attempt to describe what I do, why I do it, and how you can work within this framework.
One of the most useful things would be new hash types. I'll guide you through adding a hash type in a sample guide. Adding a hash type to the MFN Framework
Would you like to build the tools yourself? It's not terrible, but it is somewhat complicated. Build Guide
Shall we take a bone stock Ubuntu 10.04 box from zero to Cryptohaze build environment? Yes we shall! Ubuntu 10.04 Build Setup
Alternately, you can try the Ubuntu 12.04 Setup Script - much easier. In theory.
And, for the more insane, shall we take a bone stock Windows 7 box from zero to Cryptohaze? We shall try this as well! Windows 7 Build Setup
How about OS X Build Setup?
You'll need a forum account to edit the wiki. I'm trying this to reduce spambot assaults on the wiki. Once you have a forum account, please log in with that username/password and you should have edit permissions. If not, please contact me and we can debug. Thanks!