hashcat Forum

Full Version: 320.00 - faster for cudaHhashcat?
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
I got my friend to run a .bat file that I set up to see whether or not I was delusional about the increase in speed but here are his results:

https://dl.dropboxusercontent.com/u/1020...314.07.png
https://dl.dropboxusercontent.com/u/1020...320.00.PNG

The driver version is in the file name. It was also the only thing changed.

As I have a GTS 450, my numbers are close though lower.
WOW, I can verify this on my gtx560Ti.

Speeds increased drastically on 319.12 beta on Linux, too.

MD5: 1657 -> 1817 = +9.65%
SHA1: 425 -> 596 = +40.23%
SHA256: 150 -> 229 = +52.66%
SHA512: 61897 -> 70519 = +13.92%
Oracle: 69508 -> 72086 = +3.7%

All for free, I wont say no Smile

Wondering where this is coming from. Did they internally enable funnelshift for < sm_35? Its possible since its the algorithms that using rotates alot seems to gained the most speed.

Going to analyze this.
Rolf aint impressed.
GTX Titan and GTX 480.
So we can be sure now, This is about sm_21.
I knew something had to be up with the nvida vs ati crack speeds. https://hashcat.net/forum/thread-2181.html one driver update and 50-60% bump in speed. Still about 1/3 as fast but at least its a little more in proportion.