[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Date: Fri, 29 Jun 2012 01:15:44 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: sha512crypt
Claudio, Lukas -
Why doesn't this give the 11500 c/s speed on bull as mentioned before? -
$ ../run/john -te -fo=sha512crypt-opencl
OpenCL platform 0: NVIDIA CUDA, 1 device(s).
Using device 0: GeForce GTX 570
Building the kernel, this could take a while
Elapsed time: 17 seconds
Local work size (LWS) 32, global work size (GWS) 7680
Benchmarking: sha512crypt (rounds=5000) [OpenCL]... DONE
Raw: 8317 c/s real, 8287 c/s virtual
Anyhow, this is a lot better than Lukas' CUDA code speed:
$ ../run/john -te -fo=sha512crypt-cuda
Benchmarking: sha512crypt (rounds=5000) [CUDA]... DONE
Raw: 3853 c/s real, 3853 c/s virtual
Perhaps a better speed (beyond 11500 c/s) may be achieved with CUDA.
Also, luckily/surprisingly this did not "ASIC hang":
$ ../run/john -te -fo=sha512crypt-opencl -pla=1
OpenCL platform 1: AMD Accelerated Parallel Processing, 2 device(s).
Using device 0: Tahiti
Building the kernel, this could take a while
Elapsed time: 16 seconds
Local work size (LWS) 32, global work size (GWS) 16384
Benchmarking: sha512crypt (rounds=5000) [OpenCL]... DONE
Raw: 5152 c/s real, 3276K c/s virtual
For comparison, on CPU:
$ ../run/john -te -fo=sha512crypt
Benchmarking: sha512crypt (rounds=5000) [64/64]... (8xOMP) DONE
Raw: 2072 c/s real, 257 c/s virtual
(and we should be able to get almost(?) twice that by using SIMD).
Alexander
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.
