Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 24 Mar 2017 19:41:11 +0100
From: Frank Dittrich <>
Subject: Re: GPU performance

Am 24.03.2017 um 18:30 schrieb Patrick Proniewski:
> Thanks, it's way faster:
> $ ./john -test -format=raw-sha1-opencl --mask
> Device 1: GeForce GTX 1080
> Benchmarking: Raw-SHA1-opencl [SHA1 OpenCL]... Build log: 
> ptxas info    : 0 bytes gmem
> ptxas info    : Compiling entry function 'sha1' for 'sm_61'
> ptxas info    : Function properties for sha1
> ptxas         .     64 bytes stack frame, 0 bytes spill stores, 0 bytes spill loads
> ptxas info    : Used 30 registers, 16388 bytes smem, 400 bytes cmem[0], 36 bytes cmem[2]
> DONE, GPU util:100%
> Raw:	5778M c/s real, 5778M c/s virtual
> Does it mean I have to use mask mode to get this kind of performance, and other modes would run in the 87M c/s range?

Yes, you have to use --mask to get optimal OpenCL performance for very
fast hash formats.
But you don't have to use standalone mask mode, you can also use hybrid
modes, i.e., combine --mask with other modes.

While hashcat's GPU formats generate all password candidates on the GPU,
john generates the candidates on the CPU, except for very fast hash
formats which support --mask directly on the GPU.
That way, --mask avoids a data transfer bottleneck which otherwise hurts
GPU performance for those very fast hash formats.


Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.