Date: Mon, 11 Feb 2013 19:50:18 +0100 From: magnum <john.magnum@...hmail.com> To: john-dev@...ts.openwall.com Subject: Re: DCC2 performance On 11 Feb, 2013, at 16:19 , Claudio André <claudioandre.br@...il.com> wrote: > 2013/1/28 atom <atom@...hcat.net>: >> >> But for oclHashcat-plus default workload settings are a trademark >> between cracking speed and desktop lags. That is for all hashes not >> just DCC2. >> >> You can overwrite these settings and increase the workload by using >> the parameters -n and --gpu-loops > > Magnum, we can benefit from a --hash-loops option too (I measured that). > > If you create the parameter/option, each split kernel author decide to use it (or not). We already have a couple of experimental options (like --request-vector) that are silently ignored for most formats. I'm not very happy with that but I'm not sure what to do with it. We could add a bunch of flags like FMT_OPTIONALLY_VECTORIZED and FMT_TUNABLE_HASHLOOPS. Or we could continue just ignoring unsupported options. Some of my formats only support certain values for HASH_LOOPS (eg. factors of 4095). I could re-write them more like the office formats (where the final kernel do any remainder of iterations). Or we could use more abstract options like --optimize=performance and --optimize=response. I think that would be more user friendly. magnum
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.