Benchmark compare algorithms
#1
Hi everyone,

This might be a less intellegent question, in that case I am sorry Wink

I would like to know how the benchmarking option can be used to compare the algorithms
or if there is an other option to compare different algorithms?

Yes ofcourse I can see a lot of hash guesses per second and this depends on the algorithm. But this also depends on the iterations used with the algorithm if I am not mistaken. Algorithms using user defined iterations are for example bcrypt, scrypt and PBKDF2.
So I am curious which iterations values are used for the benchmark

Thank you for reading this and I hope someone can help me and clear this up
#2
(03-11-2016, 12:33 PM)raltszork Wrote: So I am curious which iterations values are used for the benchmark

The default ones, see https://hashcat.net/wiki/doku.php?id=example_hashes
#3
Oke thank you for your answer.
Is there any reason why these values were choosen?
If there is no value like SHA-512 for example, there is just 1 iteration right?
#4
Hmm, I'm not sure you fully understand what's happening here. The benchmark for 'SHA-512' is for the literal SHA-512 algorithm. If you want to see an iterated algorithm based on SHA-512, then look at the benchmarks for sha512crypt, Drupal, PBKDF2-HMAC-SHA512, etc. It's that simple. No values were "chosen"...
#5
(03-11-2016, 03:51 PM)epixoip Wrote: Hmm, I'm not sure you fully understand what's happening here. The benchmark for 'SHA-512' is for the literal SHA-512 algorithm. If you want to see an iterated algorithm based on SHA-512, then look at the benchmarks for sha512crypt, Drupal, PBKDF2-HMAC-SHA512, etc. It's that simple. No values were "chosen"...

For SHA512 you are right, no iterations.
Bcrypt does use iterations, the example uses 5 so 2^5. I was just wondering why 5. 
It doesn't really matter a lot but just wondering
#6
bcrypt is always benchmarked with a cost of 5 for historical reasons. Much easier to compare the performance of new devices to old/older devices and even different software by keeping the cost factor constant. Plus it's not hard to figure out the speed for different costs, for example to get the speed for cost of 12 just divide the benchmark speed by 2^7.
#7
5 for historical reasons ow oke, thanks a lot Smile