[3045] in cryptography@c2.net mail archive
Re: Turing Bombe story
daemon@ATHENA.MIT.EDU (Dan Todd)
Wed Jul 22 16:32:57 1998
X-Envelope-To: cryptography@c2.net
From: "Dan Todd" <dantodd@dnai.com>
To: "Clive D.W. Feather" <clive@linx.org>, <cryptography@c2.net>
Date: Wed, 22 Jul 1998 10:37:24 -0700
A couple URLs. I haven't reviewed the code so I don't know which of these
are actually decent simulations:
>From AT&T Labs UK website
Turing Bombe Model
This page contains a zip file with C++ source code which I believe to be a
reasonably reprasentative model of the Turing Bombe which was used at
Bletchley Park in WWII to find the settings for the German Enigma machine.
http://www.attlabs.att.co.uk/andyc/enigma/bombemodel.html
--
Turing Bombe Simulation
Instructions: Type the crib into the "Plain text" field and enciphered data
into the Cipher text field of the applet and press the Start button.
Alternatively use the Example button and then press Start. This should
produce a drop of 123-CBA (AH) (LY) (NS) (OR). The full test takes about 1
hour and 45 minutes on my 100 Mhz Pentium using IE 3.0. The version I have
in C (that can be downloaded from here) does the same thing in only 8
minutes.
http://www.geocities.com/CapeCanaveral/Hangar/4040/bombeapp.html
cheers,
dan
-----Original Message-----
From: Clive D.W. Feather <clive@on-the-train.demon.co.uk>
To: cryptography@c2.net <cryptography@c2.net>
Date: Wednesday, July 22, 1998 10:21 AM
Subject: Re: Turing Bombe story
>In article <3.0.3.32.19980716174533.032b5358@pop3.clark.net>, Carl Ellison
><cme@acm.org> writes
>>The brilliance of the Bombe was that it needed to test only for
>>the non-plugboard elements: (26^3)*P(5,3) = 1054560 possibilities,
>>or half that to the expected break. This comes out to about an hour
>>and a half, for a single Bombe, if it runs at 10 msec per rotor
>>position.
>
>A single Bombe unit tested one rotor order. It took about 12 minutes to
>run through the 26^3 combinations.
>
>A Bombe (the physical machine that's being rebuilt) consisted of three
>units stacked vertically, plus on some machines 4 extra wheels that were
>used for testing a possible "drop".
>
>To test a crib, 20 Bombes (60 units) were all wired up identically and set
>running. Hopefully one of the 20 would find the answer.
>
>>As for a modern Pentium beating the Bombe, maybe -- but the Bombe did an
>>incredible number of LIPS.
>
>A few weeks ago I was looking at the BP web site. Following some links, I
>fairly soon came across a Bombe simulator which I downloaded, but forgot
>where I got it from. Here's the documentation that came with it:
>
> sbgen.c sbgen.exe & sb.h
>
> This is a program that simulates how the Turing Bombe worked but
> with a few small improvements that were not present in the origional
> machine. The program sbgen is used to build a bombe program which
> is then compiled and run. It is used as follows:
>
> supergen <plain-text> <cipher-text> >super.c
> cl /O2 /Ox /Gr /G5 /Fc super.c
> super
>
> It will cycle through all 60 rotor possibilities and print out all the
> settings that it calculates are the Enigma settings along with the
> known plug board configuration. This process should not take more than
> eight minutes on a 100 Mhz Pentium.
>
> Nik Shaylor (nshaylor@tcp.co.uk)
>
>Hopefully something like AltaVista can find it for you.
>
>--
>Clive D.W. Feather | Regulation Officer, LINX | Work: <clive@linx.org>
>Tel: +44 1733 705000 | (on secondment from | Home: <cdwf@i.am>
>Fax: +44 1733 353929 | Demon Internet) | <http://i.am/davros>
>Written on my laptop; please observe the Reply-To address
>
