Checkpoint not working in latest version or me being a noob?
#1
Hey everyone,

I'm running the latest development version of Hashcat (previously was running 3.5). One thing I've noticed is I can request checkpoint quit and hashcat just happily keeps going until it's finished or until I force a quit - for example over a time span of 16 hours, requesting a checking point quit at around 10 hour mark, it'll keep going for the last 6...

Recently I threw a huge wordlist at it to the point it would take 4+ days to complete, 12 hours into the sessions I requested a checkpoint quit... 12 more hours later and it still didn't hit a restore point. Restore point sat at 0% the whole time.

Am I missing something obvious here? It always worked for me in the past - at most taking 30 minutes to hit a checkpoint... Only difference was I was using an older GTX 960, now I'm using dual GTX 1080 Ti's and latest development version. If anything I would've thought I'd hit the restore points quicker Sad

P.S.

Using WPA/WPA2 mode with anywhere from 5000 to 30,000 rules.
#2
The time it takes depends on the number of amplifiers being used, in this case the rules and also the number of non-unique salts. Both can have a large impact on the time it takes. So I guess with 30k it will take some time. Doesn't sound like a bug.
#3
Thank you so much for your response, I've done more testing and the above response is spot on. I'm literally feeding it a huge amplifier with a small password file. I've swapped them around (huge password file/small amplifier) and the checkpoints work perfectly.
#4
Right, that's basically the best way to use hashcat in every case.

If possible: For slow hashes try to hold the amplifier around 50-500 amplifier and for fast hashes 500-5000. Shift more data to the base than to the amplifier.
#5
This is just the behavior of checkpoint.

For some hash types, it can take a full day (sometimes longer) in order for a checkpoint to occur for a shutdown.

One example is the 2611 hash type (vBulletin < v3.8.5).  It has sometimes taken a full day to shut down hashcat with a checkpoint with this hash type.

I don't know which hash type is the slowest to shut down when issuing a checkpoint, but this could be worth researching.

Maybe there can be a community effort to build a chart that tracks average shutdown times for different hash types when shutting down with a checkpoint.
#6
> If possible: For slow hashes try to hold the amplifier around 50-500 amplifier and for fast hashes 500-5000. Shift more data to the base than to the amplifier.

Atom: Is this a general rule or only for those using checkpoints?
#7
This is a balance act!

The more amplifiers (rules) the longer it takes to reach a checkpoint.
The more salts you have the longer it takes to reach a checkpoint.
If you use too less amplifiers, you get a speed penalty.