+- hashcat Forum (https://hashcat.net/forum)
+-- Forum: Support (https://hashcat.net/forum/forum-3.html)
+--- Forum: hashcat (https://hashcat.net/forum/forum-45.html)
+--- Thread: hybrid attack vs. custom rules: massive performance divergence/performance optimizati (/thread-12156.html)
hybrid attack vs. custom rules: massive performance divergence/performance optimizati - fsdafsadfsdsdaf - 09-13-2024
Hey there,
this week I started experiments with hybrid attacks and maskprocessor and I am confused by the results.
I have a wordlist that itself takes about 7 minutes. I added 1 to 3 digits to every word using a) hybrid attack and b) a custom rule created with maskprocessor.
suprisingly the runtime was very different. hybrid attack took ~25 minutes (so dominated by 3*7min read time of the wordlist), the custom-rule based version was reported to take about 6 days, even though the rule engine is on gpu and should be very efficient and only read the wordlist once instead of 3 times.
If I try to append 4 digits, the custom rule even fails because 8 GB of VRAM seem to be insufficient - as a hybrid attack takes 45 minutes to append 4 digits.
questions:
1) why is the rule-based version so slow?
2) are there general rules or rules of thumb of when to use hybrid attacks and when to use custom rule
3) --increment does start 4 separate attacks after each other, which required the list to be read 3 times. Is there a way to combine these attacks to a single (hybrid attack) run?
Thank you
