MD5(Unix) bug
#1
I'm having a weird issue with the Unix MD5 hash type. After about two minutes of running, one GPU completely drops out, and the other that continues seems to not be processing correctly.

Also, when I attempt to quit using "q" after this happens, the process hangs and ultimately becomes zombied. I only seem to have an issue with this hash type, the others I've tested seem to work fine.

Thanks for the help in advance.

Code:
# /opt/oclHashcat-plus-0.09/oclHashcat-plus64.bin -m 500 -r ~/rules/generated.rule --remove --outfile test.cracked test.hash /var/passwords/wordlists/all_breaches.txt                                                                                        
oclHashcat-plus v0.09 by atom starting...

Hashes: 9 total, 9 unique salts, 9 unique digests
Bitmaps: 8 bits, 256 entries, 0x000000ff mask, 1024 bytes
Rules: 14729
Workload: 16 loops, 8 accel
Watchdog: Temperature abort trigger set to 90c
Watchdog: Temperature retain trigger set to 80c
Device #1: Cayman, 2048MB, 950Mhz, 24MCU
Device #2: Cayman, 2048MB, 950Mhz, 24MCU
Device #1: Kernel /opt/oclHashcat-plus-0.09/kernels/4098/m0500.Cayman_938.1_1.4.1741.kernel (2898852 bytes)
Device #2: Kernel /opt/oclHashcat-plus-0.09/kernels/4098/m0500.Cayman_938.1_1.4.1741.kernel (2898852 bytes)

Scanning dictionary /var/passwords/wordlists/all_breaches.txt: 1047576 bytes (0.59%), 102092 words, 150368361
Scanning dictionary /var/passwords/wordlists/all_breaches.txt: 61807285 bytes (35.03%), 6324145 words, 931483
Scanning dictionary /var/passwords/wordlists/all_breaches.txt: 136185486 bytes (77.19%), 13762913 words, 2027
Scanned dictionary /var/passwords/wordlists/all_breaches.txt: 176433682 bytes, 17781234 words, 261899766128 k
eyspace, starting attack...

[s]tatus [p]ause [r]esume [b]ypass [q]uit => s
Status.......: Running
Rules.Type...: File (/root/rules/generated.rule)
Input.Mode...: File (/var/passwords/wordlists/all_breaches.txt)
Hash.Target..: File (test.hash)
Hash.Type....: md5crypt, MD5(Unix), FreeBSD MD5, Cisco-IOS MD5
Time.Running.: 17 secs
Time.Left....: 9 days, 13 hours
Time.Util....: 17842.4ms/2036.1ms Real/CPU, 12.9% idle
Speed........:  2848.5k c/s Real,  3249.6k c/s GPU
Recovered....: 0/9 Digests, 0/9 Salts
Progress.....: 94114509/2357097895152 (0.00%)
Rejected.....: 43291341/94114509 (46.00%)
HWMon.GPU.#1.: 78% Util, 67c Temp, 30% Fan
HWMon.GPU.#2.: 78% Util, 65c Temp, 30% Fan
[s]tatus [p]ause [r]esume [b]ypass [q]uit => s
Status.......: Running
Rules.Type...: File (/root/rules/generated.rule)
Input.Mode...: File (/var/passwords/wordlists/all_breaches.txt)
Hash.Target..: File (test.hash)
Hash.Type....: md5crypt, MD5(Unix), FreeBSD MD5, Cisco-IOS MD5
Time.Running.: 1 min, 6 secs
Time.Left....: 9 days, 13 hours
Time.Util....: 66762.4ms/7571.9ms Real/CPU, 12.8% idle
Speed........:  2850.7k c/s Real,  3259.4k c/s GPU
Recovered....: 0/9 Digests, 0/9 Salts
Progress.....: 354287643/2357097895152 (0.02%)
Rejected.....: 163971099/354287643 (46.28%)
WMon.GPU.#1.: 75% Util, 77c Temp, 35% Fan
HWMon.GPU.#2.: 75% Util, 76c Temp, 33% Fan
[s]tatus [p]ause [r]esume [b]ypass [q]uit =>
Status.......: Running
Rules.Type...: File (/root/rules/generated.rule)
Input.Mode...: File (/var/passwords/wordlists/all_breaches.txt)
Hash.Target..: File (test.hash)
Hash.Type....: md5crypt, MD5(Unix), FreeBSD MD5, Cisco-IOS MD5
Time.Running.: 2 mins, 8 secs
Time.Left....: 11 days, 8 hours
Time.Util....: 128042.7ms/11064.5ms Real/CPU, 9.5% idle
Speed........:  2402.3k c/s Real,  3270.4k c/s GPU
Recovered....: 0/9 Digests, 0/9 Salts
Progress.....: 546749055/2357097895152 (0.02%)
Rejected.....: 239155839/546749055 (43.74%)
HWMon.GPU.#1.:  0% Util, 64c Temp, 57% Fan
HWMon.GPU.#2.: 99% Util, 77c Temp, 43% Fan

# ps aux | grep oclHashcat
root      1573 11.1  0.0      0     0 pts/1    Zl+  10:27   0:55 [oclHashcat-plus] <defunct>
#2
sounds like a hardware defect
#3
Downclock your cards a bit. Bad mhz/vcore voltage ratio or bad power supply.
#4
my first guess would have been that the hashing algorithm you are trying to attack is highly iterated and as a result of that, excessive load on your GPUs might cause them to hang or drop out.

second possibility is of a driver issue, make sure you are running the latest release of ATI Catalyst.

Are you able to successfully run the attacks with other slow algorithms like Wordpress, PHPBB3, WPA/WPA2 and so on.

Try to decrease the values for --gpu-loop and --gpu-accel and see if that helps.
#5
(10-14-2012, 07:34 PM)KT819GM Wrote: Downclock your cards a bit. Bad mhz/vcore voltage ratio or bad power supply.

I dropped my clocks to stock 850MHz and it looks like it's stable now. Thanks!
#6
hm, cayman running at 950Mhz... classic asic hang caused by overclocking Smile