2 ms·

Another thing you didn't catch is the difference between MD5 and the broken Windows password algorithm. And the fact that you still have to calculate MD5 for y

• dalke 14y ago
  The link and numbers I used were in regards to MD5. I even quoted "The cluster can try 180 billion combinations per second against the widely used MD5 algorithm." And no, you don't need to recompute the whole MD5 each time. Here's the Python code which shows that you can capture the hash state at an intermediate point: >>> import hashlib >>> h1 = hashlib.md5("This is the start") >>> h2 = h1.copy(); h2.update(" # Blah!"); h2.hexdigest() 'a309b70b0bc8de4e7aad1d0ac6e14b16' >>> h3 = h1.copy(); h3.update(" # Fnord!"); h3.hexdigest() 'a0a5979225e0ba46543856242daeab3c' >>> hashlib.md5("This is the start # Blah!").hexdigest() 'a309b70b0bc8de4e7aad1d0ac6e14b16' >>> hashlib.md5("This is the start # Fnord!").hexdigest() 'a0a5979225e0ba46543856242daeab3c' You may think that perhaps the copies are keeping track of the entire string. However, this is not correct. Indeed, it would make the MD5 rather useless, because it would limit processing to available memory. How would one MD5 a multi GB file with only a small amount of RAM?