Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Sat, 25 Oct 2008 22:54:53 +0400
From: SmarTeam Support <>
To: Simon Marechal <>
Subject: Re: quite fast MD5 hashing implementation

Hello Simon,

Tuesday, September 23, 2008, 5:22:46 PM, you wrote:


> It is a lot faster than everything I previously saw. I suppose he 
> achieves this speed by:
> * reversing the last round
> * working a lot on instruction/register scheduling

> Are there more clues on the russian part of the site?

One more md5 implementation:

SSE reMeDy5 is yet another single C file md5 brute forcer written by bla, with advanced features, and quite a bit of speed behind it.
It was created in an effort to speed up fast reMeDy5. It's basically the same project but implemented using SSE2 SIMD instructions. This allows x86 and x86_64 processors to execute 4 md5 calculations in parallel on each core. The calculation of 1 md5 start to finish is slower though on SSE, so do not expect to see it complete in 25% of the time of fast reMeDy5, (some tests show that it needs 50% to 70% the time fast reMeDy5 code needs).
To my knowledge this is the fastest md5 cracker of its kind. If you can write faster code, do send me a link ;-).
The single C source file(only dependency is libc) should allow you to easily drop it on any server you have laying around.
Your CPU has to support SSE2 for this code to work. To verify this type: grep sse2 /proc/cpuinfo
If your cpu doesn't, then you can still use fast reMeDy5, it's still pretty damn fast ;-).

Best regards,

To unsubscribe, e-mail and reply
to the automated confirmation request that will be sent to you.

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.