Crack

[Grincheux]

Crack is finally finished. I reorganized the database and created as many database as lengths of passwords (1 to 32).

More information on my site. In French and in english too.

Source files are joint to this post.

[Crack]
Error 0=*** Password Chloé found in 6357.243000 seconds ***
Error 1=*** Password Chloé found in 6365.231000 seconds ***

[Grincheux]

Can't use UNSIGNED INT64 even if I re-compile SQLite!

In this version I improved speed.

[Grincheux]

Improved data, strings that exist many times where defined more than once.
Changed the strings positions to improve strings aligment.
Changed PRAGMAS in SQLite
Added SQLite 3.17
Now it displays the number of combinations in the password generators.
Now "Quick" & "Find" tabs display informations when searching a password.
Created a "CrackV3.ini" file for backup and Vacuum options.
For running faster the auto-vacuum pragma is set to 0. So when inserting records there are holes. Vacuum can divide the size by 2. Even if auto-vacuum was set it would be necessaty to make a vacuum.
Added intrinsics function such as

void __stosq(unsigned long long *dst, unsigned long long val, size_t num);  (X64 only)

I made this because memset is not a macro nor an intrinsic function. I wanted to remove the call and win some cpu cycles.

[Grincheux]

The passwords I store are between 1 to 32 bytes length.
For all the tables (32) I had the same length (32) for the field [Password].
Now this field is defined with its TRUE LENGTH.

[Grincheux]

Added OpenMP.

This MD5 hash has 8CA1827022D00E01AED831634F1E77B1 been found in 281 seconds
 

[Grincheux]

Now I stop.
I make an other program.
In conclusion looking for a hash is very very long (I already knew that), and creating database for them need many work before and after.
I am pleased to have did this program because I had many problems to solve, notabily threads and speed.
In conclusion looking for cracked passwords is a loose of time.

[Grincheux]

I have a new idea.
Storing hashes in an AVL binary tree.
That would reduce the size of the database.
Hashes will be stored like in a dictionnary.
Hashes will not be stored in their full length but in many parts.
Like IBM did in DB2.

Be patient I will not be here for a long time.
The time for me to learn AVL trees and to program them.

Good night.