hashcat Forum

Full Version: Looking for help on "gpu-accel" param
You're currently viewing a stripped down version of our content. View the full version with proper formatting.

I use cudaHashcat+64, thanks for this awesome work.

I need help on the gpu-accel / workload tuning.
Help says :
Quote:scales best if you choose the number of SP per SM
NVidia = 8

What does it mean ? I have :

Quote:Platform: NVidia compatible platform found
Device #1: GeForce 8400 GS, 239MB, 1400Mhz, 1MCU
Device #1: Kernel ./kernels/4318/m1000.sm_11.64.cubin

I tried n=8, it was slow, I tried n=256 it was faster.. How many is the best ? I run the programm only overnight, so I do not need the graphic display (my screen is off).

I also saw "--gpu-loops 1000" in your posts, must I use this ? Which number ?

My use is to run it on md5/ntlm/sha1/mysql hashes overnight, with a dict and "d3ad0ne" rules.

Thanks for help.
Fastest mode is -n 320 --gpu-loops 1000/1024 .
OK, thank you !

I got :
Quote:max --gpu-loops 512 in this --mode

-m is MD5.

Why ?