Hashcat vs OclHashcat - Netntlm v1
#11
(10-07-2013, 08:26 AM)epixoip Wrote: you haven't really given us a whole lot of well-presented data

Maybe its not well presented but I think all the data you need is there. Could you specify what more data I could provide. I direct your attention to these two cases in particular:

hashcat-cli64.exe -m5500 -a0 -c 1000 -n3 --remove --pw-min=8 -o out in200 ..\Dic\04 -r append4.rule
Speed/sec.: 92.16M plains, - words

cudaHashcat-plus64.exe -m5500 -a3 --gpu-temp-disable --remove -o out in200 ?u?l?l?l?d?d?d?s
Speed.GPU.#1...: 2365.0 kH/s

(10-07-2013, 08:26 AM)epixoip Wrote: ...but you need to remember that you actually have to give gpus work to do. if you give them a small amount of work, you will not achieve full acceleration. 100MB of words is not a lot of work for fast algorithms

again.. the GPU (oclhashcat) is using a mask (?u?l?l?l?d?d?d?s) i.e. brute force NOT the 100mb word list. Is that not enough work for a GTS250?

In both case the keyspace is alpha for the first four chars and ?d?d?d?s for the char 5 to char 8. In the hashcat case this is done with the append4.rule and the 04 dictionary.

(10-07-2013, 08:26 AM)epixoip Wrote: i can't say it enough: gpus aren't magic.

epixoip didn't you just say:
(10-07-2013, 08:26 AM)epixoip Wrote: come on, everyone knows that gpus are faster than cpus! gpus are magic.

lol...sorry I just had to put that in there Tongue

(10-07-2013, 08:26 AM)epixoip Wrote: you have to optimize your attacks in order to use gpus efficiently. and in some cases, gpus are simply the wrong choice.

Yup I understand that. I am not going to crack passwords with oclhashcat 2365.0 kH/s when I can do it at 92.16M with hashcat. GPU definitely wrong choice here (thought its a bit surprising).

But my reason for starting this post was why this was the case. Because up till now I hadn't seen an example of the GPU being slower (considering I saturate it with enough words) then the CPU. And certainly not 40x slower.

Also naive as I maybe Wink you haven't given me the reason why this very specific case might slower on the GPU. I think you will agree with me this is not a case of not giving the GPU enough to work.

If required I could share test cases with you for you to check and see for your self. Thanks.
#12
your data is very poorly presented. no one is going to spend the time to try to piece together what you have and try to decipher your naming conventions and cross-reference it with the legend you provided. i'd highly suggest reading http://www.catb.org/esr/faqs/smart-questions.html

anyway,

cpu hashcat displays different speeds than oclhashcat, so you're not even talking apples and apples here. with 200 unique salts you're looking at 2.3 million words per second for plus, versus 0.46 words per second for cpu. 2.3 MH/s * 200 unique salts would be 460 million plains per second. so in this case it looks like your gpu is about 5x faster than your cpu.
#13
(10-07-2013, 08:26 AM)epixoip Wrote: EDIT: oh, one other thing you need to realize as well is that brute force with cpu hashcat is always going to underperform due to the way generated candidates are fed to the worker threads. cpu hashcat is best suited for wordlist-based attacks.

Cool! Thats interesting. Nice info. Thanks!

(10-07-2013, 08:26 AM)epixoip Wrote: I really think that your biggest problem is misunderstanding your tools... what their strengths/weaknesses are, when to use which, how to maximize their efficiency...

hashcat's Strength: using dic+rules over oclhashcat (mask) passwords of these types
?a?a?a?a?a?a?a?a

Really I get it. You seem to have understanding of the strengths/weaknesses of these tools because of experience from using them. And you seem to except that all as been done to optimize the given tool.

I brought this up because I felt this was an area that oclhashcat might possible be optimized to get higher speed on NETNTLMv1 hash.
#14
(10-07-2013, 09:38 AM)epixoip Wrote: your data is very poorly presented. no one is going to spend the time to try to piece together what you have and try to decipher your naming conventions and cross-reference it with the legend you provided. i'd highly suggest reading http://www.catb.org/esr/faqs/smart-questions.html

Ouch! ok point noted.

(10-07-2013, 09:38 AM)epixoip Wrote: cpu hashcat displays different speeds than oclhashcat, so you're not even talking apples and apples here. with 200 unique salts you're looking at 2.3 million words per second for plus, versus 0.46 words per second for cpu. 2.3 MH/s * 200 unique salts would be 460 million plains per second. so in this case it looks like your gpu is about 5x faster than your cpu.

Lol I didn't know that! Just saw this: "One last note about performance. There was a change to the status-display of the speed value which does not affect the real performance. With new oclHashcat-plus v0.15 the speed that is shown gets divided by the number of uncracked unique salts. Older oclHashcat-plus versions did not take this into account. Don't get shocked when you're cracking a large salted hashlist and the speed dropped by hundret of times (or to be exact by number of hashes/salts), the total time will stay equal."

This makes sense. Thank you epixoip!