Disclaimer
MDCrack is a security tool
designed to attack various hash algorithms at a very fast rate.
This tool is provided to assist pen-testers, system administrators and users in testing their own passwords, or other passwords for which they
have been given permission to test, in order to determine the security of such passwords.
it is not intended to assist with breaking local and remote hosts and accounts.
I hereby disclaim any responsibility for illegal actions taken based on this software.
This software is by no mean a virus/trojan/malware as suggested by some anti-virus products like Mcafee that labels it as potentially unwanted.
However, note that if you were to find a MDCrack process unexpectedly running on your server like it has been reported by some people, then your system has likely been compromised and is being exploited.
In such a case, please do not blame MDCrack but rather fix your security holes.
Also, take the time to check the file MD5 signature available on this site to avoid using a tampered version.
Warranty and limitations of liability
SOFTWARES AND DOCUMENTATIONS ARE PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT.
IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE,
ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Well to make it brief, use it at your own risk and please, in a legal manner
About MDCrack
MDCrack is a free featureful password cracker designed to bruteforce several commonly used hash algorithms at a very aggressive speed rate.
It can retrieve any password made of up to 16 characters and allowing up to 55 characters with an additionnal user salt.
In order to achieve the highest possible speed rate, this program uses several cores for each algorithm it supports.
Each one of these cores provides a different level of optimization dynamically selected at run time to best adapt with changing candidates length.
To date, this program supports bruteforce attacks for 21 algorithms: MD2, MD4, MD5, HMAC-MD4, HMAC-MD5, FreeBSD, Apache, NTLMv1, IOS and PIX (both enable and user) hashes,
Invision Power Board 2.x (IPB2), MD4MD4, MD4MD4S, MD5MD5, MD5MD5S, PHP, PHPS, CRC32, CRC32B, ADLER32 the list of algorithms is growing up.
Multi threading allows for parallel cracking, distributing workload between available CPUs and thus multiplying overall speed by the number of processor(s) used.
This yields to phenomenal performance with the advent of multi-cores CPUs design and even on the old generation of multi-CPUs systems.
Several options offer fine tuning for the attack and allow to save and resume your session at a later time.
This program has multiple uses in real life, while its primary goal is obviously to test the strength of hashed passwords,
it can also be used as a raw benchmark tool for integer calculs.
For any question you may have, please be sure to first read the Frequenty Ask Questions file (FAQ) provided in MDCrack ZIP file.
I hope you will enjoy this program.
Latest News
Apr 29 2007.
Version 1.8(3).
That was quite a long time but here it is.
This minor update includes following new features and modifications:
|
Improved Win32 support
|
- MDCrack will try to register the .mds file extension with its own executable name and icon
- Default session filename has been renamed "latest.mds", thus double clicking on it will resume the latest.mds saved session
- Note that the .mds suffix won't be automatically set to your session filename, either specify it at the command line or rename your old session files
|
|
Bug fixes
|
This new release comes with its usual truck load of bug fixes.
As always your bug reports are very appreciated and useful to make this program better.
Feel free to send them at mdcrack@videotron.ca
Thank you.
|
|
All cores (all algorithms) upgraded
|
All cores have been improved to support up to 16 characters candidates (55 if adding a custom salt set with the --prepend and --append options).
Also default maximal candidate size has been changed accordingly, use the --maxsize and --minsize options to override with your own limits if necessary.
|
|
CRC32 CRC32B ADLER32 cores added
|
While these hash algorithms are not really related to passwords security, i thought they could prove useful in some situations
and would be, at the very least, interesting to implement and evaluate.
They can most likely be optimized deeper but i won't spend too much time on this.
If you wish to share your work on any of these algorithms, please contact me.
Also note that because they produce very short in size hashes (32bits) you might well collide before testing the expected candidate.
In this situation, use this mysterious --all command line option and MDCrack will simly resume its work after each new found collision.
Well, now there is a justification for this option :)
|
|
Cores loader design improved
|
All cores can be dynamically loaded at runtime allowing MDCrack to choose the best optimized core for a given candidate.
Using this approach, all currently supported algorithms can be further optimized.
For now, only the CRC32 CRC32B and ADLER algorithms fully exploit this new design while all other algorithms will eventually follow
but gradually. (you can try "mdcrack-sse --algorithm=CRC32 --test-hash" as an example)
|
Feb 25 2007.
Version 1.8(2).
*Nix users may want to know that MDCrack has been reported working fine with Wine for Linux (reported by N.Spohrer)
More information on Wine can be found here
This update includes following new features and modifications:
|
Zip package
|
- Lr4.exe decoder has been added to the MDCrack package
- Faq has been updated to reflect some of the changes, still very incomplete though (lack of time)
Some good souls around willing to help ?
|
|
Bug fixes
|
- Closing the console at runtime won't trigger a system popup anymore
- Session related command line options (session, resume, delete, autosave etc...) were sensitive to their order of appearance
in the command line. This could lead to inconsistent and somewhat confusing results for some command line configurations.
As for now, "session=FILE" option globally set the default session filename for the whole command line while it can be overridden
locally for each option (if a FILE argument is available).
- Accuracy for algorithms auto-detection feature has been improved (can be tested with --guess option)
- For sessions using a color scheme, a new console is spawned thus leaving the original one untouched and clean.
- One could not add a comma to the custom charset string since comma was also the suboptions separator.
There is no more suboptions separator, predefined character subsets can be inserted anywhere along the charset string
using the %x metacharacter notation where x denotes a specific character subset.
Just repeat % twice (%%) to specify a % character.
Currently available predefined character subsets are:
%U (all upercase alpha) %L (all lowercase alpha) %P (all punctuation) %N (all numericals) %C (all control characters)
- Several memory leaks have been fixed
- Other few minor bug fixes...
|
|
Non printable characters
|
Non printable characters can be easily used within custom charsets and salts using a #nn notation, where the 'nn' part designates the hexadecimal
value of the character. Use a double '##' to specify a # character alone.
Try --charset=%C as an example.
|
|
Syslog / output / detach options
|
MDCrack now supports builtin redirection of its output to a file or a syslog server for local and remote monitoring.
Also a --detach option allows for closing the console without terminating the process.
The builtin syslog client offers ability to setup syslog severity filter, syslog facility, server address, protocol and port number.
|
|
Runtime commands
|
Several keyboard shortcuts allow control and monitoring at runtime
- Esc: Shortcuts summary
- F[1-5]: Change process priority, F1 for IDLE ... F5 for HIGH
- F10: Autosave on/off switch
- F11: Change color scheme
- F12: Fullscreen on/off switch
- Enter: Show process information
- Space: Show thread(s) statistics
- Ctrl-c: Interrupt current session
|
Dec 30 2006.
New companion tool, Lr4.
|
Tool added.
|
Lr4.exe is a small console based companion tool for decoding of Loginrecovery V4 dump files.
Hash entries dumped from the floppy disk (version 04) are shadowed and unusable with free tools such as MDCrack and John.
This tool decodes and formats LR dump files in the familiar Pwdump style thus recovering your informations for offline cracking with your favorite tool.
Also its small size allows it to conveniently live on the same loginrecovery floppy disk.
See Lr4.exe --help for a list of all available command line options.
|
Dec 17 2006.
Version 1.8(1).
This minor update includes the following modifications:
|
Benchmark
|
The process of benchmarking has been speeded up due to the removal of the tests using --prepend and --append salt options.
However all natively salted algorithms (eg: FreeBSD) are obviously still salted during the process but are now tested only once.
This change was made because the use of --prepend and --append options were not significantly impacting the performances.
If you wish to test each algorithm several time in a row for better accuracy, use the PASS argument in --benchmark=PASS.
|
|
New PHP cores
|
Two cores have been added to support generic "PHP" algorithm based upon a MD5(hex(MD5(pass))) function.
It is basically the same algorithm as MD5MD5 except that the inner hash is hex encoded before being hashed.
Use --algorithm=PHP at command line.
|
|
New PHPS cores
|
Two cores have been added to support generic "PHPS" algorithm based upon a MD5(hex(MD5(pass)).salt) function.
Same as MD5MD5S with a hex encoded inner hash.
Use --algorithm=PHPS at command line.
|
|
Bug fix for IPB2 cores
|
One password out of 2 couldn't be correctly found because of a small bug in these two cores.
This issue has been fixed and IPB2 cores have been also optimised a bit further. (~ +6% faster on a 1xP4 1.66)
|
Dec 10 2006.
Version 1.8.
|
Some few changes
|
Session file format has been reworked:
Verbosity has been removed and must now be manually reset when resuming a session.
eg: mdcrack-sse --resume --verbose=very
Added: a salt field to support new algorithms (MD5MD5S MD4MD4S IPB2).
Added: option --show to show content of a session file without resuming it.
Several new cores have been written, see below for details.
Please do not forget to refresh the performance table by sending your benchmarks.
|
|
New MD5MD5 cores
|
Two cores have been added to support generic "MD5MD5" algorithm based upon a MD5(MD5(pass)) function.
Use --algorithm=MD5MD5 at command line.
|
|
New MD4MD4 cores
|
Two cores have been added to support generic "MD4MD4" algorithm based upon a MD4(MD4(pass)) function.
Use --algorithm=MD4MD4 at command line.
|
|
New MD5MD5S cores
|
Two cores have been added to support generic "MD5MD5S" algorithm based upon a MD5(MD5(pass).salt) function.
These cores need a salt that can be provided through the --salt command line option.
Use --algorithm=MD5MD5S at command line.
|
|
New MD4MD4S cores
|
Two cores have been added to support "MD4MD4S" algorithm based upon a MD4(MD4(pass).salt) function.
These cores need a salt that can be provided through the --salt command line option.
Use --algorithm=MD4MD4S at command line.
|
|
New IPB2 cores
|
Two cores have been added to support Invision power board hashing algorithm based upon a MD5(hex(MD5(salt)).hex(MD5(pass))) function.
These cores need a salt that can be provided through the --salt command line option.
Use --algorithm=IPB2 at command line.
|
Nov 26 2006.
Version 1.7.5 is a maintenance release including several bug fixes and some new features.
What's new in 1.7.5 ?
|
Bug fixes
|
Following issues have been reported by users.
- Benchmarking with 10 passes per algorithm would eventually crash MDCrack because of a heap corruption.
- When launched with multiple threads, MDCrack was ignoring the requested minimum size for candidates and was starting a session with candidate size set to 1 character.
- Benchmark priority has been reduced to NORMAL by default in order to make the process more responsive to interruptions.
- Some other minor problems in MDCrack and Cis7
These issues have been fixed.
|
|
Features added
|
Following features have been added.
- MDCrack now tries to auto-detect algorithm of hash provided from the command line, this means you can avoid using --algorithm option in much cases
- MDCrack's command line interface has been improved with the use of long options and sub-options for better clarity, FAQs have also been modified to reflect the changes.
- Users's custom charsets can now be easily written using built-in character subsets (eg. %L for all lowercase alphanumerics), see option --help for a complete list of subsets.
- Option '--test-hash' will use one of MDCrack's built-in hash instead of a user defined hash (the same built-in hash used during a benchmark). That allows easy testing of the various available options.
- Option '--autosave' will periodically save session during runtime thus protecting all the hard work from a system failure
- Cis7 command line interface has also been upgraded with long options. A new '--file' option can import and decode ciphertexts directly from an IOS configuration file
|
Oct 30 2006.
Version 1.7 is available.
What's new in 1.7 ?
|
Several bug fixes
|
Following issues, related to the handling of saved sessions, have been identified.
- Under scenarios involving the use of a prefix salt, saved last candidates were being corrupted
and saved sessions were irrelevant afterward.
- A similar bug, specific to NTLM1 algorithm, was corrupting the saved maximum candidate size.
- HMAC based sessions couldn't be interrupted and later resumed since their nonces were not being saved at all.
These issues have been fixed.
|
|
FreeBSD
Apache
Cisco IOS
Cores
|
6 new cores have been added in order to support FreeBSD MD5, Apache MD5 and Cisco IOS algorithms.
All of them are based on FreeBSD implementation of MD5, involving a salt and ~1000 MD5 iterations.
So do not expect millions of hash per second,
at least till a network-distributed version of mdcrack has been released
Send me your benchmarks here to refresh the performance table.
|
|
Tool added to
MDCrack Zip file
|
Cis7.exe is a small satellite tool for en/decoding of Cisco passwords type 7.
Passwords type 7 are an implementation of Vigenere used by Cisco IOS to protect configuration passwords from unwise eyes.
This algorithm is known to be weak for more than a decade now and despite the much more robust
alternative proposed by Cisco and based on FreeBSD MD5, it is not uncommon to see router passwords stored in this format.
More info at cisco.com
The pletoria of tools i have been able to test failed to correctly decode large passwords or passwords encrypted with a large seed (eg: 4578702849070453)
and so i wrote mine.
You can download Cis7.exe here, use "Cis7 -h" on the command line to get a little bit more info...
|
Oct 18 2006.
Version 1.6 is available.
Below is a change summary for version 1.6:
|
Several bug fixes
|
- Some errors have been accidentally inserted in MD4 and MD5 cores during the process of rewriting the old version 1.2
These errors could, for instance, prevent MDCrack from correctly finding a MD4/MD5 collision if used with a prefix salt.
- Another bug prevented MDCrack from selecting some MD5 cores 3 leading to some sessions unexpectedly aborting.
All these issues are now fixed.
Please report any other problem you may find here
|
|
New HMAC Cores
|
8 new cores have been added in order to support HMAC-MD4 and HMAC-MD5 algorithms.
Use option -D to specify a HMAC message/nonce.
Send me your benchmarks here to refresh the performance table.
|
Sept 08 2006.
Faq updated.
Sept 06 2006.
It has been a long time (actually years)
since the 1.2 version, too long for all the exciting work still in pending state. So MDCrack is back from the dead.
Also these few web pages were getting old and needed a refresh.
For the sake of clarity and simplicity, all information related to this project is now accessible from this single web page.
Most of the old web site content has gone away however you can still download part of the old stuffs here.
As ever this software is free, donations are now accepted and very much appreciated.
As ever your suggestions, bug reports and benchmarks are also very much appreciated.
About the new 1.5 version:
This new version is based on a complete rewrite of the now old 1.2, it
includes a lot of new features along with an impressive performance
gain versus the 1.2.
At this time 1.5 only supports Win32 platforms, that includes Win XP,
Win2K and probably Win NT. (The later being untested, please report any
problem you may have here)
Vista should be ok but has not been tested yet, if you test it please send me an update. (confirmed by A.Masalskis, see perf table)
A version for Linux and FreeBSD might be done in the future but is currently not on the top of the stack.
Below is a summary of all the great new features for this 1.5 version:
|
New Package format
|
The new release is provided as a zip file made of a FAQ and
a binary optimized with Intel SIMD extensions, MMX SSE and SSE2,
but nevertheless compatible with older processors through an auto-detect feature.
|
|
New command line options among which...
|
- Charset random generation
- Charset random mix up
- Candidate maximum size limit
- Candidate comparison count limit
- User defined first candidate
- User defined session filename
- Console color schemes
- Console full screen
- and several other options...
|
|
Detailed statistics at runtime
|
Several people have
complained about the lack of information provided during a session.
Indeed, except for the size of tested candidates, nothing was
reported during a session runtime. This new release allows to get
threads statistics at runtime by simply pressing a key, these statistics are very
verbose but do not penalize cores performance (or marginally). As a
result you now have a mean to know exactly, thread by thread, which
candidate is being tested, which core is being used, the average speed,
an estimation of the remaining time before keyspace depletion etc...
|
|
better support for Win32
|
As you have surely noticed
it, the console interface of the latest (and first) win32 release was a
bit rustic not to say hostile. It was basically a raw port from the
original Unix based 1.2 version to Windows. This new version best
supports Microsoft operating systems with a set of options to configure process priority,
process console, saved sessions and multi-threading.
|
|
New benchmarks interface
|
The process of benchmarking all cores one by one could be somewhat boring.
It has been automated and improved to get the best of all algorithms from a single command line.
Because MDCrack is very CPU intensive these benchmarks can also be used, in a more general way,
to rawly evaluate and compare CPU(s) speed for integer based calcul.
|
|
Support for MultiCPUs
|
MDCrack now makes use of Windows threads implementation.
It supports parallel cracking and load sharing between up to 8 CPU/threads and, de facto, increase your overall speed performance by as
many CPU(s) you wish to provide for your session.
Note that MDCrack lets you fix the total number of threads and also offers support for
automatic threads mask affinity that glues each one of the threads to a dedicated CPU.
A first stone to the still to come network distributed version.
|
|
New NTLM1 Cores
|
4 new cores have been added to get optimal speed from different candidate/salt scenarios.
This translates for instance by a 100% speed gain for 9 characters candidates on a Pentium IV versus 1.2 version.
See the performance table below for details.
These cores can be used to test Microsoft passwords used in the old fashion (but still in use) authentication scheme NTLMv1.
Note: I will most probably write some cores for HMAC MD5 and its unicode variant (NTLMv2) one day or another, check this page frequently for an update.
|
|
New PIX Cores
|
2 new Cores dedicated to test
Cisco PIX Firewalls passwords for both Enable and Users accounts.
(work for all versions including latest 7.x images)
Note: PIX Users accounts are used to grant access to the firewall itself or the services it provides like VPN and web access and can therefore be
interesting targets for an attacker seeking to compromise your private network or to bypass your security policy.
|
|
New MD2 Cores
|
4 new Cores to test MD2 hashed passwords.
There might still be work to be done in order to get optimal speed performance for these 4 cores.
Althought MD2 was originally designed and optimized for 8 bits architectures, i may have overlooked something on the way and frankly i hope so.
If someone has some great ideas in regard to MD2 optimization on a 32 bits architecture, please contact me
|
How does it compares with previous releases and at which speed rate can i expect to run it on my system ? see performance table
Downloads
Current version is 1.8(3)
Before downloading, be sure to read and understand the legal notice
Always take care to look at the provided MD5 signatures (below) preferably on this site prior to execute MDCrack, especially if you download it from elsewhere.
If you don't already have a tool to proceed, you may want to try this free one from http://www.fourmilab.ch/.
If you want to redistribute this release, please also provide these MD5 signatures or at least a link to it.
MDCrack 1.8(3) -- Win32 -- Zip file (Contains FAQs and binaries)
MD5 signature -- b401be875677b43792902f4f72253674 mdcrack-183.zip
MD5 signature -- c678f3defc8462c64d77efc7b52a113d mdcrack-sse.exe
MDCrack 1.83 -- FAQ 1.8 (TXT)
MDCrack 1.83 -- FAQ 1.8 (HTML)
Lr4 1.0 (also included in MDCrack zip file)
MD5 signature -- 05be81c4a5997367242f197e4d1bc694 lr4.exe
Cis7 1.2 (also included in MDCrack Zip file)
MD5 signature -- ec22a038dccc806281d9d233f1683ddd cis7.exe
Performance
Following table is mostly updated with your benchmarks; If your configuration doesn't already appear in this list, please email me yours.
If you have a multi-CPUs system or any modern CPU like Intel core 2 duo and AMD X2 at hands,
i would be very interested to get a copy of your benchmarks.
Other benchmarks based on older single CPUs are also welcomed but preferably not too old (PIII or above).
Do not hesitate to contact me if you think an error has been introduced in this table.
Since version 1.7.5, use "mdcrack-sse --benchmark" on the command line to produce a benchmark.
When it starts the tests, MDCrack automatically creates as many threads it finds CPUs in your system with a default NORMAL process priority in order to keep responsiveness for interruptions.
It would be wise to close all other greedy applications before starting the process to achieve best result.
Note that even if the program will let you change several default parameters such as timer, process priority and the number of threads from the command line;
You are discouraged to proceed so with benchmarks you send because it may lead to sub-optimal speed reports.
Just wait the few minutes that takes, copy/paste and send your email.. that's all
Thank you.
This table reports speed performance by version and system configuration for each algorithm best core (core 1 when available).
For those who may wonder, cores speed is expressed in hashes per second (h/s), this is the same as comparisons per second or candidates per second (c/s).
Each hash is computed from a candidate before being compared with the target hash therefore all three aforementioned measuring units are identical for MDCrack.
The same remark does not apply for every password cracking software, some of them accept several targets in the same process and thus may perform several comparison with a given candidate hash.
In order to avoid confusion and get the real picture, speed performance is expressed here in hashes per second.
PIX Enable and PIX User Cores give almost identical benchmarks and this is no surprise because Cisco hash format for PIX users is a salted version of Enable hash
with some few other modifications though.
The real surprise comes from MD2 benchmarks where Intel CPUs seem to outperform AMD.
Hopefully some optimization may still be possible but, meantime, let's wait for more benchmarks from more recent CPUs ...
|
MDCrack
|
CPU
|
System
|
Period (s)
|
MD2 (h/s)
|
MD5 (h/s)
|
MD4 (h/s)
|
NTLM1 (h/s)
|
PIX Enable (h/s)
|
PIX User (h/s)
|
HMAC-MD5 (h/s)
|
HMAC-MD4 (h/s)
|
FreeBSD (h/s)
|
IOS (h/s)
|
Apache (h/s)
|
Binary / Compiler
|
|
1.7
|
2x XEON 3.2GHz (DC + HT)
|
Win XP Pro
|
5
|
1 884 043
|
42 299 451
|
68 475 279
|
61 791 052
|
31 790 823
|
29 875 043
|
3 823 043
|
5 737 982
|
13 816
|
14 104
|
13 887
|
mdcrack-sse
|
|
1.7
|
1x AMD X2 4200+ 2.28Ghz (DC)
|
WinXP Pro SP2
|
5
|
432 274
|
20 875 320
|
34 731 624
|
32 792 140
|
15 402 937
|
15 408 374
|
2 876 551
|
4 333 737
|
11 618
|
11 636
|
11 477
|
mdcrack-sse
|
|
1.7
|
1x P4 2Ghz (DC)
|
WinXP Pro
|
5
|
444 013
|
19 963 045
|
33 289 080
|
31 584 267
|
15 175 891
|
15 166 281
|
2 995 789
|
4 552 167
|
12 270
|
12 551
|
12 298
|
mdcrack-sse
|
|
1.7
|
1x P4 2.41GHz
|
Win XP
|
5
|
433 371
|
7 037 468
|
10 641 954
|
9 610 838
|
5 190 657
|
5 181 395
|
856 343
|
1 293 849
|
3 765
|
3 956
|
3 748
|
mdcrack-sse
|
|
1.5
|
2x XEON 3.2GHz (DC + HT)
|
Win XP Pro
|
5
|
1 884 043
|
42 299 451
|
68 475 279
|
61 791 052
|
31 790 823
|
29 875 043
|
|
|
|
|
|
mdcrack-sse
|
|
1.5
|
2x XEON 3.6Ghz (HT)
|
Win 2k
|
5
|
1 265 337
|
26 855 210
|
41 117 627
|
39 397 498
|
18 365 742
|
19 039 391
|
|
|
|
|
|
mdcrack-sse
|
|
1.5
|
1x AMD X2 4800+ 2.42Ghz (DC)
|
Win XP
|
5
|
476 736
|
22 995 292
|
38 880 061
|
36 206 314
|
17 037 139
|
17 081 129
|
|
|
|
|
|
mdcrack-sse
|
|
1.5
|
1x P4 1.66Ghz (T2300) (DC)
|
Win Vista RC1
|
5
|
389 990
|
14 414 805
|
23 179 434
|
22 074 546
|
10 514 795
|
10 425 183
|
|
|
|
|
|
mdcrack-sse
|
|
1.5
|
1x P4 3Ghz (HT)
|
Win XP
|
5
|
453 196
|
11 021 967
|
17 243 695
|
16 447 991
|
8 180 054
|
8 180 109
|
|
|
|
|
|
mdcrack-sse
|
|
1.5
|
1x AMD XP-mobile 2.32Ghz
|
Win XP
|
5
|
228 974
|
10 090 138
|
16 093 819
|
15 884 568
|
7 103 820
|
7 104 434
|
|
|
|
|
|
mdcrack-sse
|
|
1.5
|
1x AMD64 3000+ 2Ghz
|
Win XP
|
5
|
195 786
|
9 397 021
|
15 766 544
|
15 062 335
|
6 968 033
|
6 977 514
|
|
|
|
|
|
mdcrack-sse
|
|
1.5
|
1x P4 2.66Ghz
|
Win XP
|
5
|
490 688
|
8 724 678
|
11 219 466
|
9 690 632
|
6 802 660
|
6 160 632
|
|
|
|
|
|
mdcrack-sse
|
|
1.5
|
2x P3 1Ghz
|
Win 2k
|
5
|
233 623
|
7 989 055
|
12 366 112
|
11 937 637
|
5 672 357
|
5 675 329
|
|
|
|
|
|
mdcrack-sse
|
|
1.2
|
Athlon 1.8 Ghz
|
Win XP Pro
|
27
|
|
5 808 949
|
9 334 967
|
5 489 462
|
|
|
|
|
|
|
|
Visual C++
|
|
1.2
|
Athlon 1.53 Ghz
|
Win XP
|
10
|
|
5 080 455
|
8 503 911
|
4 800 505
|
|
|
|
|
|
|
|
Visual C++
|
* DC for dual core, HT for hyper-threading
Contacts and Acknowledgment
Author: Gregory Duchemin ( mdcrack@videotron.ca )
I would like to thank the following people for their work and kindness:
Simeon Pilgrim for his great work on MD5.
Solar Designer, author of John the ripper, for kindly providing MDCrack with a sweet home in Moscow.
Christophe Braud who took for him the burden to port version 1.2 to win32.
Karma, Dark Neo and NullAdd for sharing their frontends.
All the people who have made donation, sent bug reports, benchmarks and ideas...
Thank you!
Old stuff
Before downloading, be sure to read and understand the legal notice
FreeBSD and Linux Gentoo users, you may already have MDCrack 1.2 in your CDs, check in the ports list.
Sources in these ports have most probably been patched to best fit in your operating system, i would recommend to try them first.
Always take care to look at the provided MD5 signatures, preferably on this site, prior to use sources and binaries.
Below are links to sources and binaries for the previous 1.2 version.
Links
Also take the time to visit these sites...
From the same author,
DHCPing is a security tool for auditing DHCP servers.
Openwall GNU/*/Linux (or Owl for short),
A security-enhanced operating system with Linux and GNU software as its core...
Dataforce is a russian Internet provider,
they have hosted MDCrack for years and for free.
Packetstormsecurity,
a very large collection of security stuff. A great site.
