Posts: 47
Threads: 15
Joined: Dec 2013
Today, I did a little experiment between hashcat 3.10 and oclhashcat 2.01.
This test was decided upon because I had been noticing of late how long
hashcat 3.10 was taking to complete a specific task.
I ran a bruteforce on 961 md5 hashes and the results not only surprised me
but backed up my theory.
In the first test I ran it for 4 hours and it reached a completion of 8.21%. Using the
same hcmask and computer, I ran the same file again using a previous version of
hashcat, and it took less than 3 hours to surpass the same target. 8.35% complete.
Can someone explain this phenomenon?
Posts: 2,267
Threads: 16
Joined: Feb 2013
02-13-2017, 05:17 PM
(This post was last modified: 02-13-2017, 05:17 PM by philsmd.)
This seems to be a very, very, very vage problem description.
We do not even know what your setup looks like etc.
Also, why would you test with hashcat 3.10 and not with the latest version?
I don't understand
Posts: 47
Threads: 15
Joined: Dec 2013
Firstly, regardless of the setup the difference in speed is noticeable ( it was tested on the exact same setup as
I only have one computer set up to do this kind of thing (Windows 10 os just case you were wondering .)
Secondly, I have been using 3.10 for sometime now and found it stable and as it is an update to previous
versions and I believe it is supposed to be faster than all prior versions if I remember reading correctly. ( This apprantly
is not the case)
As mentioned in my original post nothing was changed and more importantly no one was hurt in the making of this observation, it is just a general query as to why this could happen
Posts: 2,267
Threads: 16
Joined: Feb 2013
02-13-2017, 05:39 PM
(This post was last modified: 02-13-2017, 05:39 PM by philsmd.)
yeah, the answer is that the focus is to always support the newest and fastest GPUs and that's why there is sometimes a trade-off/compromise that might impact older systems.
That's why I ask for the setup (including make and model of your GPU(s)).
The problem with answering to problem reports affecting older versions of hashcat is that we do not really fix/support/maintain older versions of hashcat... devs always improve the latest version.
That's why we always tell users to use the latest version and double-check if the problems still exist with the newest version.
Posts: 47
Threads: 15
Joined: Dec 2013
I understand what you are saying. I decided on the test as I was considering updating my graphics cards to the NVidia 1080 based on the slower speeds I have experienced with 3.10 and with the age of my current ones which are 5900 series and I believe I read somewhere that there is a slight problem with the opencl.
I have checked what driver version I am using and it falls into the correct category as stated on the homepage.
I will give the newer version a try and see if that makes a difference if not will revert to the old one until I update my cards.
Posts: 5,185
Threads: 230
Joined: Apr 2010
59xx series is dead, even AMD dropped it from the drivers.
Posts: 47
Threads: 15
Joined: Dec 2013
Aye, I ken they are now defunct, wasn't aware that they had been dropped from the drivers (using 15.201.1151.1008) but they are very much alive in my system and still crunching the numbers lol.
Will do till I decide to replace to NVidia.
Posts: 5,185
Threads: 230
Joined: Apr 2010
OK, the 15.201.1151.1008 is an older driver version. Unfortionally we can't stick to an older driver version because we need to support new GPU and they need drivers..
