hashcat Forum
statsprocessor only running on one core - Printable Version

+- hashcat Forum (https://hashcat.net/forum)
+-- Forum: Support (https://hashcat.net/forum/forum-3.html)
+--- Forum: hashcat-utils, maskprocessor, statsprocessor, md5stress, wikistrip (https://hashcat.net/forum/forum-28.html)
+--- Thread: statsprocessor only running on one core (/thread-2171.html)



statsprocessor only running on one core - jpenna - 03-26-2013

Hi, I'm starting to learn how to use your great set of applications, and i have a question:

is the threat title behaviour the expected behaviour?
Runnig sp64.bin (v0.08) on The-Distribution-Which-Does-Not-Handle-OpenCL-Well (Kali) Linux

Thanks


RE: statsprocessor only running on one core - epixoip - 03-26-2013

statsprocessor is single threaded.


RE: statsprocessor only running on one core - jpenna - 03-26-2013

Hey, thanks epixoip, the wiki description "(per core)" misled me.

now i'm trying:

running -- time sp64.bin (options) > /dev/nul = 16min (100% 1core use)

running -- sp64.bin (same options) | cudahascat (benchmark nv recommended tuning) = ~102.7M/s
(sp64.bin~50% 1core use) = 36min

Is there anything that i am missing? md5 hashcat-benchmark.sh = 1444.6M/s


RE: statsprocessor only running on one core - epixoip - 03-26-2013

first, you don't need to pipe statsprocessor into cudaHashcat-plus, as statsprocessor is built-in. all mask-based attacks use markov mode by default unless you explicitly disable it.

second, of course it will be slower piping it into the program. since you're piping, the plains are streamed to the gpu in real time, instead of being preloaded into a buffer like with wordlist-based attacks or generated on the gpu like mask attack. this, plus the overhead of computing the algorithm, accounts for all of your overhead.