Comments on Distracted: SHAbr update, I passed 60 Mhashes/s

tnx for that, i might give it a try some day... if...

2011-09-22T23:51:25.791+02:00

tnx for that, i might give it a try some day... if i ever get back to some coding fun...
(even too lazy to log on :P)

Here's a faster SSE_Endian_Reverse32. It shoul...

2011-09-18T22:16:23.523+02:00

Here's a faster SSE_Endian_Reverse32. It should compile to 6 instructions including moves vs 12 instructions including moves. With AVX it's 5 vs 9 and AVX with XOP it's 4 vs 5.

#define ROT16(n) _mm_shufflelo_epi16(_mm_shufflehi_epi16(n, 0xb1), 0xb1)
#define SSE_Endian_Reverse32(a) _mm_or_si128(_mm_srli_epi16(ROT16(n), 8), _mm_slli_epi16(ROT16(n), 8))

Hi Daniel, this is great tool,but maybe you can h...

2011-02-17T02:42:09.527+01:00

Hi Daniel,
this is great tool,but maybe you can help me with compiler error
1>.\sha1.cpp(35) : error C3820: 'm': initializers must be managed
1> Aligned data types not supported in managed code
I use vs 2008 and couldn't find solution for this
Thanks

I neglected to mention that I first tried with 1 a...

2010-03-23T15:16:15.138+01:00

I neglected to mention that I first tried with 1 and 4 'a' characters too. The results were never as expected. I was also checking for results which were in the wrong byte order etc but no luck. Maybe I'm doing something wrong but I need help to figure out what! :-)

"I took the code and fed it a 16 byte key wit...

2010-03-22T09:40:33.069+01:00

"I took the code and fed it a 16 byte key with all 'a' characters (the other 3 keys I left as zeroes) but the resulting sha1 didn't match the sha1 generated by a known to be working but normal slow speed sha1 routine. Should it match or have I misunderstood something about how the password cracker works? Thanks, Simon"

if i remember correctly, shabr is optimized for passwords with a length of max 15, so 16 results in a wrong hash yes

"hello, i search precomputed raibowtable for ...

2010-03-22T09:39:42.086+01:00

"hello, i search precomputed raibowtable for sha1, in your website you say, all is mooved at garr, but the folder for sha1 is empty :("

tables are currently being synchronized to garr, ntlm tables are being transferred at the moment

I took the code and fed it a 16 byte key with all ...

2010-03-21T21:24:54.473+01:00

I took the code and fed it a 16 byte key with all 'a' characters (the other 3 keys I left as zeroes) but the resulting sha1 didn't match the sha1 generated by a known to be working but normal slow speed sha1 routine. Should it match or have I misunderstood something about how the password cracker works? Thanks, Simon

hello, i search precomputed raibowtable for sha1, ...

2010-03-21T19:59:47.051+01:00

hello, i search precomputed raibowtable for sha1, in your website you say, all is mooved at garr, but the folder for sha1 is empty :(

Daniel, There is a small syntax error at sha1.cpp:...

2010-03-15T00:10:16.963+01:00

Daniel,
There is a small syntax error at sha1.cpp:19

- #ifdef WIN32
+ #ifdef _WIN32

Just as your comment says, windows is stupid. :)

Still Rampant,
Durandal

Hi Daniel, thanks for this wonderfull tool. i wan...

2009-11-30T17:33:15.345+01:00

Hi Daniel,
thanks for this wonderfull tool. i wanted to modify it to use it for my needs. I have an SHA1 hash and also the data which is been hashed but i miss the salt. I wanted to use your source to try to find the salt of this hash. the data which is hashed is known so this could make bruteforce easyier. The data is 6 digits long but the salt is unknown. The salt could also be other chars non standard ASCII. As i am not into CPP programming what tool have you used to compile SHABR? and if you could help me to set up the program so proper to meet my needs.
wbr and good luck

hi there, seems like you just want to always set &...

2009-10-03T13:20:46.308+02:00

hi there, seems like you just want to always set 'length' to 16. Is that correct? Or is it always 2*length ?

0x80 is used to set a bit to '1' at the end of the string, it is then padded with 0's.

This is a great implementation. Great work! Howeve...

2009-10-03T11:09:42.020+02:00

This is a great implementation. Great work! However, I need to modify this a little bit, and I am not able to do that. Let me explain what the problem is.

If I encrypt string "12345678" (8 bytes, in hexa 31 32 33 34 35 36 37 38) you get a hash 7c222f... (etc.). Your algorithm finds the original text in a few seconds. However, I need to crack the same string, but encrypted a little bit different. Instead of 8 bytes, it was 16, with the last 8 as 0 (null terminator). In other words it was (in hexa) 31 32 33 34 35 36 37 38 00 00 00 00 00 00 00 00. Of course, that produced another hash value and your algoritm does not find 12345678 in this case.

I tried modifying different things but nothing worked. But then my knowledge of the SHA1 algoritm are very limited and don't think I can modify it. For instance I didn't understand the meaning for this:

// padding bytes
pMuteMe1[length] = 0x80;
pMuteMe2[length] = 0x80;
pMuteMe3[length] = 0x80;
pMuteMe4[length] = 0x80;

Could you suggest the changes required for modifying this brute force algorithm for my purposes?

Thank you.

hi tnx for that, i might have a look at it one day...

2009-07-18T00:40:33.842+02:00

hi tnx for that, i might have a look at it one day :)

i kind of fixed a linux version once already, main problem there was something with the timers to calculate the speed when using multiple threads. was a while ago, i might probably fix that fairly fast by now :p

if i'm not mistaken i just used _mm_store_si128 or something to get data from sse registers to memory. not sure anymore :) shabr would actually need a fair bit of adjustment, implementing many optimizations from my other brute forcers. I guess it should be doable to get it passed 100 Mhashes/s on my system.

I made some changes to the code to make it work in...

2009-07-18T00:22:30.478+02:00

I made some changes to the code to make it work in OS X and (probably) linux/freebsd/etc.

http://s3.amazonaws.com:80/burkelibbey.baconfile.com/shabr-unix.tgz

This might require SSE4.1. I'm not quite sure. I had to use __builtin_ia32_vec_ext_v4si.

That was it! Thanks.

2009-05-26T12:28:31.350+02:00

That was it! Thanks.

You might need to install: http://www.microsoft.c...

2009-05-25T20:13:46.208+02:00

You might need to install:

http://www.microsoft.com/downloads/details.aspx?familyid=D5692CE4-ADAD-4000-ABFE-64628A267EF0&displaylang=en

I need anything else to run this? When I execute i...

2009-05-25T17:44:46.251+02:00

I need anything else to run this? When I execute it it gives and error.

hi there, you might want to take a look at www.cry...

2009-05-03T12:35:00.000+02:00

hi there, you might want to take a look at www.cryptohaze.com, this guy has just release a new beta version of his cuda brute forcer, with sha-1 support, and per position charsets:

http://www.cryptohaze.com/forum/viewtopic.php?f=4&t=57&start=0

haven't tested it myself yet, but it looks like this might suit your needs. He will release source code later on as well, so that might be even easier to alter if you still need it a little different.

Cool, I stumbled upon this because.. well.. I fckd...

2009-05-02T16:40:00.000+02:00

Cool, I stumbled upon this because.. well.. I fckd up...
Anyway, my password consists of certain sequences that I'm sure of, just not the order of them (and if all of them are available) also there is some *glue* of normal characters between the sequences. What I want to do is simply add a couple of new "characters" (that actually consists of x characters).
So that "after z" comes a new character which really is "abcdefg" or soemthing.

With this I think I can get the password down to about 8-9 "charactes" which would be very doable.

Now I should be able to alter the source to do this myself (really appretiate you sharing the code) however my C is a bit rusty/nonexistent so before banging my head in the wall I just wonder if this would be as easy to add as it sounds in my head or perhaps if someone knows about an application that has this functionality from the start.