hashcat Forum

With the most recent oclHashcat (2.01) and AMD GPUs every time I try and run a brute force attack (-a3) against a WPA hash (algorithm 2500) ocl will lock up. You can press the status key once and it will print out each GPU is running at 0 H/s. After you press status key once you do not get the prompt back.

Dictionary attacks work just fine. 

kill and kill -9 of the oclHashcat64.bin process does not do anything. The machine has to be rebooted to recover from the failure.

Its not really a critical bug for me, brute force attack typically yields nothing useful for WPA. However I figured you may want to know about the issue.

Details:
Debian kernel 3.16.0-4
fglrx 15.20.3
oclHashcat 2.01

Command line used:
oclHashcat64.bin -a3 -m2500 hash.hccap

It's called an ASIC hang. Search the forums.

You were correct it was an ASIC hang. Most of the searches I saw pointed to a hardware fault of some kind, either GPU is bad or power supply. However, my issue seems to be software related. I removed all the GPUs from my chassis, leaving only one. If I run oclHashcat64 as a normal user it works fine with a dictionary attack. Once a normal (non-root) user tries to run a mask attack the ASIC hang occurs. However, if the root user runs the mask attack, everything works fine.

Can anyone shed any light on this issue? I would rather not have to run hashcat as root.

(01-29-2016, 09:21 PM)epixoip Wrote: [ -> ]It's called an ASIC hang. Search the forums.