[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20120501053903.GA10619@openwall.com>
Date: Tue, 1 May 2012 09:39:03 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: Lukas - status report #2
Lukas -
So I implemented OpenMP in your wpapsk_postprocess() - please see the
attached patch. The new speeds are (still without tuning):
user@...l:~/john/magnum-jumbo/src$ ../run/john -te -fo=wpapsk-cuda
Benchmarking: wpapsk-cuda [GPU]... (8xOMP) DONE
Raw: 32385 c/s real, 16490 c/s virtual
80% of hashcat's - but we also use CPU, which hashcat probably does not
(or not to the same extent).
user@...l:~/john/magnum-jumbo/src$ ../run/john -te -fo=wpapsk-opencl -pla=1
OpenCL platform 1: AMD Accelerated Parallel Processing, 2 device(s).
Using device 0: Tahiti
Optimal Group work Size = 64
Benchmarking: wpapsk-opencl [GPU - OpenCL]... (8xOMP) DONE
Raw: 55138 c/s real, 41090 c/s virtual
user@...l:~/john/magnum-jumbo/src$ ../run/john -te -fo=wpapsk-opencl -pla=1 -dev=1
OpenCL platform 1: AMD Accelerated Parallel Processing, 2 device(s).
Using device 1: AMD FX(tm)-8120 Eight-Core Processor
Optimal Group work Size = 2
Benchmarking: wpapsk-opencl [GPU - OpenCL]... (8xOMP) DONE
Raw: 2123 c/s real, 266 c/s virtual
For context, non-OpenMP speeds were:
On Tue, May 01, 2012 at 09:10:46AM +0400, Solar Designer wrote:
> user@...l:~/john/magnum-jumbo/run$ ./john -te -fo=wpapsk-cuda
> Benchmarking: wpapsk-cuda [GPU]... DONE
> Raw: 27015 c/s real, 27151 c/s virtual
>
> user@...l:~/john/magnum-jumbo/run$ ./john -te -fo=wpapsk-opencl
> OpenCL platform 0: NVIDIA CUDA, 1 device(s).
> Using device 0: GeForce GTX 570
> OpenCL error (CL_OUT_OF_RESOURCES) in file (opencl_wpapsk_fmt.c) at line (125) - (Run kernel)
>
> user@...l:~/john/magnum-jumbo/run$ ./john -te -fo=wpapsk-opencl -pla=1
> OpenCL platform 1: AMD Accelerated Parallel Processing, 2 device(s).
> Using device 0: Tahiti
> Optimal Group work Size = 64
> Benchmarking: wpapsk-opencl [GPU - OpenCL]... DONE
> Raw: 40069 c/s real, 100800 c/s virtual
>
> For comparison, this used to be "64000 c/s real, 531692 c/s virtual".
> So the real speed has decreased, and the CPU usage has increased.
>
> Did you say that the old speeds were somehow not real? Was there some
> kind of bug (incomplete implementation?) that you fixed?
>
> user@...l:~/john/magnum-jumbo/run$ ./john -te -fo=wpapsk-opencl -pla=1 -dev=1
> OpenCL platform 1: AMD Accelerated Parallel Processing, 2 device(s).
> Using device 1: AMD FX(tm)-8120 Eight-Core Processor
> Optimal Group work Size = 2
> Benchmarking: wpapsk-opencl [GPU - OpenCL]... DONE
> Raw: 2094 c/s real, 266 c/s virtual
Alexander
View attachment "wpapsk-gpu+openmp.diff" of type "text/plain" (2279 bytes)
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.
