Benchmark vs Mask NTLM
#1
Hi everyone,

I tried searching the forum for some guidance on the difference (if any) and could not come up with anything definitive. 

Basically I'm trying to understand the significant difference I'm getting between the benchmark and a password mask for ?a?a?a?a?a?a?a?a 

The benchmark suggests 60,837MH/s (66GH?/s) but when performing actual work it's only doing ~30,000 MH/s .  Is this inline with typical real world performance or is something seriously busted :/

Using 4.2.1

Benchmark:
Hashmode: 1000 - NTLM                                                                                                                                  
Speed.Dev.#1.....: 32759.2 MH/s (60.68ms) @ Accel:256 Loops:512 Thr:1024 Vec:2
Speed.Dev.#2.....: 33077.8 MH/s (60.06ms) @ Accel:256 Loops:512 Thr:1024 Vec:2
Speed.Dev.#*.....: 65837.0 MH/s

Mask Attack
Command: hashcat -m 1000 -O -a 3 -w 4 output.ntds ?a?a?a?a?a?a?a?a

Guess.Mask.......: ?a?a?a?a?a?a?a?a [8]                                                                                                                                                                   
Guess.Queue......: 1/1 (100.00%)                                                                                                                                                                              
Speed.Dev.#1.....: 15437.0 MH/s (249.24ms) @ Accel:256 Loops:1024 Thr:1024 Vec:2                                                                                                                              
Speed.Dev.#2.....: 15656.7 MH/s (245.88ms) @ Accel:256 Loops:1024 Thr:1024 Vec:2                                                                                                                              
Speed.Dev.#*.....: 31093.8 MH/s                      
                                  

HWMon.Dev.#1.....: Temp: 82c Fan: 59% Util:100% Core:1582MHz Mem:3802MHz Bus:16
HWMon.Dev.#2.....: Temp: 81c Fan: 58% Util:100% Core:1683MHz Mem:3802MHz Bus:8
#2
tsk0 Wrote:Hi everyone,

I tried searching the forum for some guidance on the difference (if any) and could not come up with anything definitive. 

Basically I'm trying to understand the significant difference I'm getting between the benchmark and a password mask for ?a?a?a?a?a?a?a?a 

The benchmark suggests 60,837MH/s (66GH?/s) but when performing actual work it's only doing ~30,000 MH/s .  Is this inline with typical real world performance or is something seriously busted :/

Using 4.2.1

Benchmark:
Hashmode: 1000 - NTLM                                                                                                                                  
Speed.Dev.#1.....: 32759.2 MH/s (60.68ms) @ Accel:256 Loops:512 Thr:1024 Vec:2
Speed.Dev.#2.....: 33077.8 MH/s (60.06ms) @ Accel:256 Loops:512 Thr:1024 Vec:2
Speed.Dev.#*.....: 65837.0 MH/s

Mask Attack
Command: hashcat -m 1000 -O -a 3 -w 4 output.ntds ?a?a?a?a?a?a?a?a

Guess.Mask.......: ?a?a?a?a?a?a?a?a [8]                                                                                                                                                                   
Guess.Queue......: 1/1 (100.00%)                                                                                                                                                                              
Speed.Dev.#1.....: 15437.0 MH/s (249.24ms) @ Accel:256 Loops:1024 Thr:1024 Vec:2                                                                                                                              
Speed.Dev.#2.....: 15656.7 MH/s (245.88ms) @ Accel:256 Loops:1024 Thr:1024 Vec:2                                                                                                                              
Speed.Dev.#*.....: 31093.8 MH/s                      
                                  

HWMon.Dev.#1.....: Temp: 82c Fan: 59% Util:100% Core:1582MHz Mem:3802MHz Bus:16
HWMon.Dev.#2.....: Temp: 81c Fan: 58% Util:100% Core:1683MHz Mem:3802MHz Bus:8

Trying to find info on overclocking and found this post; so yeap...it's expected (multihash).

https://hashcat.net/forum/thread-5051-po...l#pid27999
#3
Your cards are overheating. Find a way to cool them better.
#4
The issue here is you're trying to compare single hash performance to multi-hash performance. Apples vs. oranges.
#5
epixoip Wrote:The issue here is you're trying to compare single hash performance to multi-hash performance. Apples vs. oranges.

Hi epixoip,

Yeap, you said the same in that other thread I linked above .  Do you think it's worthwhile putting that info on the wiki or FAQ somewhere.

Regards
TSK0
#6
Sorry, didn't see that reply. Kind of got lost in the wall of quoted text. In the future, please do not quote other people's posts in your reply unless doing so adds significant value to your reply. And even then, please only quote the relevant portion. Like this:

(10-09-2018, 07:46 AM)tsk0 Wrote: Do you think it's worthwhile putting that info on the wiki or FAQ somewhere.

It kind of is covered on the FAQ: https://hashcat.net/faq#why_is_my_attack_so_slow