Open impakho opened 7 years ago
Hello, did Pyrit show any error (traceback)? also, please post the contents of ~/.pyrit/config
and pyrit list_cores
thanks
Machine A:
root@nano:~# cat .pyrit/config
default_storage = file://
limit_ncpus = -1
rpc_announce = true
rpc_announce_broadcast = false
rpc_knownclients =
rpc_server = false
use_CUDA = false
use_OpenCL = true
workunit_size = 75000
root@nano:~# pyrit list_cores
Pyrit 0.5.1 (C) 2008-2011 Lukas Lueg - 2015 John Mora
https://github.com/JPaulMora/Pyrit
This code is distributed under the GNU General Public License v3+
CPUs disabled in config...
The following OpenCL GPUs seem aviable...
#1: 'OpenCL-Device 'Tahiti''
root@nano:~#
Machine B:
root@mini:~# cat .pyrit/config
default_storage = file://
limit_ncpus = -1
rpc_announce = true
rpc_announce_broadcast = false
rpc_knownclients = 192.168.1.124
rpc_server = true
use_CUDA = false
use_OpenCL = true
workunit_size = 75000
root@mini:~# pyrit list_cores
Pyrit 0.5.1 (C) 2008-2011 Lukas Lueg - 2015 John Mora
https://github.com/JPaulMora/Pyrit
This code is distributed under the GNU General Public License v3+
CPUs disabled in config...
#1: 'Network-Clients'
The following OpenCL GPUs seem aviable...
#1: 'OpenCL-Device 'Tahiti''
root@mini:~#
No traceback or any error appear. When I run an example GPU crack on Machine B, it tells me something like "Tried 4091 PMKs so far; 935 PMKs per second". It seems like a normal situation. But several seconds later, the counting was just stopped. Is it a thread blocking?
I see, well that's weird. still I'll see what I can do, I'm currently considering a rewrite of CUDA & OpenCL modules since they were made long ago, but that will probably take me a while. If you'd like, watch this repo so you get notified with new updates.
Hello! Looks like i have same problem. I use pyrit 0.5.1, with CUDA.
When i configure pyrit to use only GPU(via limit_ncpus
) and use batch
command- process run smoothly. But when i attack .pcap with attack_db
command and disabled CPU, it hangs at start with something like:
Tried 374941 PMKs so far (3.9%); 268777394 PMKs per second
and in strace i see:
futex(0x555f783cab90, FUTEX_WAIT_BITSET_PRIVATE|FUTEX_CLOCK_REALTIME, 0, NULL, 0xffffffff
If i enable CPU, attack_db
command works fine.
Ok, I think I know whats happening. db lookup happens on the CPU so blocking CPU computing globally makes it hang. I think I can push a fix later today.
run cpu crack, it was a success, the password was came out. run gpu(opencl) crack, it will automatically stop. run cpu+gpu(opencl) crack(it's the fastest), it was okay, i got the password too.
When I run gpu crack, I also run strace and debug it. Here is the output info from strace:
We can see in strace hang on FUTEX_WAIT_PRIVATE.