Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Sat, 4 Jul 2015 18:51:14 +0200
From: Agnieszka Bielec <bielecagnieszka8@...il.com>
To: john-dev@...ts.openwall.com
Subject: Re: PHC: Lyra2 on GPU

2015-07-04 18:11 GMT+02:00 Solar Designer <solar@...nwall.com>:
> On Sat, Jul 04, 2015 at 05:55:16PM +0200, Agnieszka Bielec wrote:
>> [a@...er run]$ ./john --incremental=digits --min-length=8
>> --max-length=8 hash --format=lyra2-opencl
>> Loaded 1 password hash (Lyra2-opencl, Lyra2 [Lyra2 Sponge OpenCL
>> (inefficient, development use only)])
>> Device 0: Tahiti [AMD Radeon HD 7900 Series]
>> Local worksize (LWS) 64, global worksize (GWS) 1024
>> Device 0: Tahiti [AMD Radeon HD 7900 Series]
>> Local worksize (LWS) 64, global worksize (GWS) 1024
>
> Why are these two lines repeated?

because I'm using reset(), not init()

>
>> Press 'q' or Ctrl-C to abort, almost any other key for status
>> 0g 0:00:00:03 0.03% (ETA: 21:57:45) 0g/s 8268p/s 8268c/s 8268C/s
>> GPU:48°C util:64% fan:20% 12302145..10052501
>>
>>
>> [a@...er run]$ ./john --incremental=digits --min-length=8
>> --max-length=8 hash --format=lyra2-opencl --dev=5
>> Loaded 1 password hash (Lyra2-opencl, Lyra2 [Lyra2 Sponge OpenCL
>> (inefficient, development use only)])
>> Device 5: GeForce GTX TITAN
>> Local worksize (LWS) 64, global worksize (GWS) 2048
>> Device 5: GeForce GTX TITAN
>> Local worksize (LWS) 64, global worksize (GWS) 2048
>> Press 'q' or Ctrl-C to abort, almost any other key for status
>> 0g 0:00:00:08 0.05% (ETA: 23:14:10) 0g/s 5905p/s 5905c/s 5905C/s
>> GPU:60°C util:99% fan:38% 19777123..11312252
>>
>>
>> and CUDA
>>
>>
>> [a@...er run]$ ./john --incremental=digits --min-length=8
>> --max-length=8 hash --format=lyra2-cuda
>> Loaded 1 password hash (Lyra2-cuda, Lyra2 [Lyra2 CUDA])
>> /
>> Press 'q' or Ctrl-C to abort, almost any other key for status
>> 0g 0:00:00:02 0.00% (ETA: 08:26:18) 0g/s 1432p/s 1432c/s 1432C/s
>> GPU:49°C util:88% fan:34% 02082424..01223312
>
> Can you let these run, say, for 1 minute each?

I was waiting around 1 minute but speed didn't change much

but now I'm including precise speed (maybe I was waiting more than 1 min)

--dev=1
0g 0:00:02:07 1.05% (ETA: 22:53:05) 0g/s 8273p/s 8273c/s 8273C/s
GPU:70°C util:84% fan:36% 03641302..09089964

--dev=5
0g 0:00:02:43 0.98% (ETA: 00:05:28) 0g/s 5985p/s 5985c/s 5985C/s
GPU:80°C util:99% fan:58% 12080924..12214428

cuda
0g 0:00:02:25 0.28% (ETA: 09:54:25) 0g/s 1939p/s 1939c/s 1939C/s
GPU:76°C util:96% fan:51% 10250143..18012230

sorry but previously in cuda version blocks and threads wasn't set to
optimal numbers
I have no idea why on AMD real password cracking gives better speed

> Also, in a separate test, please actually let them crack the password
> after a while - e.g., use a wordlist and put the correct password around
> line 100000.

line 97k, and one hash is in file 'hash'

[a@...er run]$ rm john.log john.pot
[a@...er run]$ ./john hash --format=lyra2-opencl --wordlist=italian
Loaded 1 password hash (Lyra2-opencl, Lyra2 [Lyra2 Sponge OpenCL
(inefficient, development use only)])
Device 0: Tahiti [AMD Radeon HD 7900 Series]
Local worksize (LWS) 64, global worksize (GWS) 1024
Device 0: Tahiti [AMD Radeon HD 7900 Series]
Local worksize (LWS) 64, global worksize (GWS) 1024
Press 'q' or Ctrl-C to abort, almost any other key for status
password         (?)
1g 0:00:00:09 DONE (2015-07-04 19:48) 0.1015g/s 8212p/s 8212c/s
8212C/s GPU:55°C util:84% fan:21% disfagia..dissipassero
Use the "--show" option to display all of the cracked passwords reliably
Session completed

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.