[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20120624195937.GC5319@openwall.com>
Date: Sun, 24 Jun 2012 23:59:37 +0400
From: Solar Designer <solar@...nwall.com>
To: john-dev@...ts.openwall.com
Subject: Re: xsha512-opencl, raw-sha512-opencl on HD 7970
On Sun, Jun 24, 2012 at 04:46:11PM +0800, myrice wrote:
> I fixed the bug. It is due to I incidentally comment memcpy in
> set_key() and commit to magnum-jumbo. Now it is:
>
> [12:43:22 myrice] run $ ./john -te=1 -fo=xsha512-opencl -pla=1
> OpenCL platform 1: AMD Accelerated Parallel Processing, 2 device(s).
> Using device 0: Tahiti
> Compilation log: LOOP UNROLL: pragma unroll (line 163)
> Unrolled as requested!
> LOOP UNROLL: pragma unroll (line 168)
> Unrolled as requested!
> LOOP UNROLL: pragma unroll (line 183)
> Unrolled as requested!
>
> Optimal Group work Size = 128
> Global work size = 2097152
> Benchmarking: Mac OS X 10.7+ salted SHA-512 [OpenCL]... DONE
> Many salts: 54176K c/s real, 1083M c/s virtual
> Only one salt: 26214K c/s real, 41943K c/s virtual
Looks good. I think it should be possible to improve the "many salts"
speed to about 150M c/s, though. ;-) Here's a relevant tweet by
@gat3way (Milen):
Well anyway....152M/s singlehash SHA512 on 7970 pic.twitter.com/tWmGn28U
Thanks,
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.
