Date: Thu, 28 Jun 2012 18:39:05 -0300 From: Claudio André <claudioandre.br@...il.com> To: john-dev@...ts.openwall.com Subject: Re: sha512crypt Em 28-06-2012 18:15, Solar Designer escreveu: > Claudio, Lukas - > > Why doesn't this give the 11500 c/s speed on bull as mentioned before? - You have to select better than my safe defaults. I recommend: LWS=512 GWS=7680 ../run/john -te -fo=sha512crypt-opencl BTW: sane defaults are a good improvement, but, at least on AMD, it is a little bit tricky (i mean, there are a lot of good LWS/GWS that hang 7970, using the same source code). > Also, luckily/surprisingly this did not "ASIC hang": I can hang 7970 if i try to, but, i haven't seen it for a while. And here (when i shake the GPU), i haven't seen for long time. > $ ../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 You can double CPU speed using SIMD? Something i would like to see. Claudio
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.