Benchmarks and Password per second
#1
Hi everyone,

I should collect some data on the performance of the latest version of hashcat.
In particular, I'm interested at computing the number of passwords that can be hashed in the time unit for different hash functions. 
At the moment, I'm using the option -b (e.g., ./hashcat -m 17600 -b), obtaining the number of hashes per second (e.g., 1893.2 MH/s).

Now, it is not clear to me if the MH/s means "million of passwords hashed per second" or  "million of hash iterations per second". If the second interpretation is the correct one, I can compute the number of passwords hashed per second by dividing it for the number of rounds of the hash function, right?

Thanks,
Reply
#2
(kilo, mega, giga, tera) hashes per second

several algorithms do not have iterations (raw hashes for instance). The iteration count is mentioned separately, just try it yourself:
Code:
hashcat -m 7400 -b
Reply
#3
(04-24-2020, 08:41 AM)philsmd Wrote: (kilo, mega, giga, tera) hashes per second

several algorithms do not have iterations (raw hashes for instance). The iteration count is mentioned separately, just try it yourself:
Code:
hashcat -m 7400 -b

Ok, thanks. Let me check if I understood correctly. 
Here, some examples:

In case of bcrypt:
Code:
Hashmode: 3200 - bcrypt $2*$, Blowfish (Unix) (Iterations: 32)

Speed.#1.........:    60227 H/s (41.42ms) @ Accel:16 Loops:8 Thr:8 Vec:1
I have that the number of passwords per second is 60227 / 32 (where, 32 is the number of Iterations).

Instead, for SHA3-512:
Code:
Hashmode: 17600 - SHA3-512

Speed.#1.........:  1894.5 MH/s (54.59ms) @ Accel:64 Loops:32 Thr:640 Vec:1
1894.5 MH/s is the actual number of passwords per second, as no number of iterations is reported anywhere in the output.

Is this correct?
Reply
#4
(04-24-2020, 10:59 AM)theNosieKnows Wrote:
Code:
Hashmode: 3200 - bcrypt $2*$, Blowfish (Unix) (Iterations: 32)

Speed.#1.........:    60227 H/s (41.42ms) @ Accel:16 Loops:8 Thr:8 Vec:1
I have that the number of passwords per second is 60227 / 32 (where, 32 is the number of Iterations).

no, the number of guesses per second is 60227 for a bcrypt hash with 32 iterations
Reply