oclhash question about length
#1
Hi,

Maybe a stupid question, but, as written in the oclhashcat documentation :

Code:
Charset: 0123456789
Length: 8
oclHashcat.exe hashes.txt -m 0 -o res.txt ?d?d?d?d ?d?d?d?d

If I need Length 9 which is the best/faster :
oclHashcat.exe hashes.txt -m 0 -o res.txt ?d?d?d?d ?d?d?d?d?d
or
oclHashcat.exe hashes.txt -m 0 -o res.txt ?d?d?d?d?d ?d?d?d?d
?

Thank you.
#2
The first.
Left mask works better if it's divisible by 4.
Right is less dependant.
However, I may be wrong.
Maybe ?d?d?d?d?d?d ?d?d?d is better.
How about finding out?
#3
if left side is of length 4, 8 or 12 its most performant. there is one exception if you get this warning:

WARNING: words in dict_left < .... Can't gain full performance

in this case its better to put more chars on the left side.

but its like rolf said, just find it out by running it and compare performance.
#4
I did the benchmark.
Each run executed 3 times, then AVG of the time is taken.

?d?d?d?d ?d?d?d?d?d : 4.3363
?d?d?d?d?d ?d?d?d?d : 3.3016
?d?d?d?d?d?d ?d?d?d : 3.2763
?d?d?d?d?d?d?d ?d?d : 3.6256
#5
Thank you for your answers.
It works better with 4 at left you're right.

Another question :
the mask ?d?d?d?d will crack only password of 4 digits isn't it ?
According to my test it will not crack the password '123'.

How can I specify, like in hashcat, the lenght ?
Code:
--bf-pw-min 1 --bf-pw-max 9

If it is not possible, must I have to write something like :

Code:
oclHashcat.exe hashes.txt -m 0 -o res.txt ?d
oclHashcat.exe hashes.txt -m 0 -o res.txt ?d ?d
oclHashcat.exe hashes.txt -m 0 -o res.txt ?d?d ?d
oclHashcat.exe hashes.txt -m 0 -o res.txt ?d?d ?d?d
...
in order to crack all digits from 1 to 9 length ?
Thank you
#6
use --increment, see http://ob-security.info/?p=116 too. my recommendation: do not try to crack passes < length 4 using oclHashcat
#7
Nice feature.
Can you please confirm that this command

Code:
oclHashcat.exe hashes.txt -m 0 --increment ?d ?d?d?d?d?d?d?d?d?d

will try length from 2 (?d ?d) to 10 ?

But with this, I'm not sure to be as fast as what we write before..
#8
It will, but it's gonna be slow as a 200 years old grandma.
I suggest using:
Code:
?d ?d
?d?d ?d
?d?d?d ?d
--increment ?d?d?d?d ?d?d?d?d?d
#9
Hi,

During this night I tried with 10 hashes from length 1 to 10 :
Code:
oclHashcat.exe hash.txt -m 0  ?d?d?d?d?d ?d?d?d?d?d

It only cracks 2 hashes : the one with length 6 and the one with length 10. Strange.

Then I tried with --increment :
Code:
oclHashcat.exe hash.txt -m 0  --increment ?d?d?d?d?d ?d?d?d?d?d

It cracks me 7/10 hashes : ones from length 4 to 10.

@Rolf
Here is my/your code, on 10 hashes too :
Code:
oclHashcat.exe tmp.txt -m 0 ?d ?d
oclHashcat.exe tmp.txt -m 0 ?d?d ?d
oclHashcat.exe tmp.txt -m 0 ?d?d?d ?d
oclHashcat.exe tmp.txt -m 0 --increment ?d?d?d?d ?d?d?d?d?d

First, I had many warnings like "WARNING: words in dict_left < .... Can't gain full performance"

It has been completed in less than 1 minute, so faster than before !!

It also cracks 8/10 hashes : ones from length 2 to 9. (not hashes with length 1 nor 10 ; len 1 might be not possible)
#10
something must be wrong with your test, i did my own and everything works. here is each step:

root@sf:~/oclHashcat-0.24# cat ints
1
11
111
1111
11111
111111
1111111
11111111
111111111
1111111111

then i md5'ed them:

root@sf:~/oclHashcat-0.24# cat hash
c4ca4238a0b923820dcc509a6f75849b
6512bd43d9caa6e02c990b0a82652dca
698d51a19d8a121ce581499d7b701668
b59c67bf196a4758191e42f76670ceba
b0baee9d279d34fa1dfd71aadb908c3f
96e79218965eb72c92a549dd5a330112
7fa8282ad93047a4d6fe6111c93b308a
1bbd886460827015e5d605ed44252251
bbb8aae57c104cda40c93843ad5e6db8
e11170b8cbd2d74102651cb967fa28e5

and then i let run oclhashcat:

root@sf:~/oclHashcat-0.24# ./oclHashcat64.bin -m 0 hash -d 1 -n 8 1 111111111 --increment
oclHashcat v0.24 starting...

Digests: 10 entries, 10 unique
Bitmaps: 8 bits, 256 entries, 0x000000ff mask, 1024 bytes
Platforms: 1
Platform #1: Advanced Micro Devices, Inc., OpenCL 1.1 ATI-Stream-v2.3 (451) (2 matched)
Device #1: Juniper, 256MB, 0Mhz, 10MCU
Device #1: Kernel kernels/4098/m0000.64.Juniper.kernel (105725 bytes)
Device #2: Juniper, 256MB, 0Mhz, 10MCU
Device #2: skipped by user
WARNING: words in dict_left < 5120. Can't gain full performance
WARNING: words in dict_right < 256. Can't gain full performance
6512bd43d9caa6e02c990b0a82652dca:11
698d51a19d8a121ce581499d7b701668:111
b59c67bf196a4758191e42f76670ceba:1111
b0baee9d279d34fa1dfd71aadb908c3f:11111
96e79218965eb72c92a549dd5a330112:111111
7fa8282ad93047a4d6fe6111c93b308a:1111111
1bbd886460827015e5d605ed44252251:11111111
bbb8aae57c104cda40c93843ad5e6db8:111111111
e11170b8cbd2d74102651cb967fa28e5:1111111111
[s]tatus [p]ause [r]esume [h]elp [q]uit =>
Threads...: 1
Mode.Left.: Mask '1' (1)
Mode.Right: Mask '111111111' (1)
Speed.GPU1: 1/s (finished)
Speed.GPU*: 0/s
Recovered.: 9/10 Digests, 0/1 Salts
Progress..: 1/1 (100.00%)
Running...: 0 secs
Estimated.: 0 secs

the first cannot not be cracked because it has the size 1 and oclhashcat needs at least size 2 because of one char per side.