hashcat Forum

Full Version: Keepass performance
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Hi, I have recently been comparing hashcat benchmarks and noticed a big performance dip between earlier versions of hashcat (5.x.x or 4.x.x) and the newer version (6.x.x) when it comes to Keepass cracking. 

In this benchmark a GeForce RTX 2080 Ti with hashcat 5.0.0 achives about 218200 H/s per GPU.
https://gist.github.com/binary1985/c8153...157562763e

In this benchmark a GeForce RTX 2080 Ti with hashcat 4.2.0 achives about 224650 H/s per GPU.
https://gist.github.com/codeandsec/1c1f2...b061944675

In this benchmark a GeForce RTX 2080 Ti with hashcat 6.0.0 achives about 63292 H/s per GPU.
https://gist.github.com/OhGodANoob/69545...e661212422

I have only compared the performance of MD5, SHA-1, SHA2-256, NTLM, and NetNTLMv2, and Keepass. The difference in performance with Keepass stands out because not even the newest GPUs (6900 XT or RTX 3090) can get close to the past benchmarks.
3090: https://gist.github.com/Chick3nman/e4fce...2106d73fef
6900 XT: https://hashcat.net/forum/thread-9854.html
As pointed out by jsteube on Github, the reason there _seems_ to be a dip in performance is that newer benchmarks (6.x.x) use a 4x iteration count for KeePass, therefore the hashrate is lower.