+- hashcat Forum (https://hashcat.net/forum)
+-- Forum: Misc (https://hashcat.net/forum/forum-15.html)
+--- Forum: Hardware (https://hashcat.net/forum/forum-13.html)
+--- Thread: GTX 970 Benchmark (/thread-3982.html)
Pages:
1
2
GTX 970 Benchmark - TheJoSko - 01-16-2015
Greetings everyone.
Long time lurker and just built my own rig with the new NVidia cards.
I went with two GTX 970's as they are readily available, in my budget, power efficient, and easily found in reference design.
I haven't seen any benchmarks around for these cards, so here is mine.
https://gist.github.com/joshuaskorich/7678d4fcbf49acb69387
I feel it is rather impressive for NVidia hardware and furthers a lot of the claims that these new cards are pretty sweet.
Setup is Ubuntu 14.04x64 and I'm using the official NVidia 343.22 drivers and cudaHashcat-1.31.
xorg.conf includes:
Code:
Section "Device"
Option "Coolbits" "12"
Option "RegistryDwords" "PerfLevelSrc=0x2222"
Option "Interactive" "False"
EndSectionAt the moment I am very mildly overclocking with (note: cannot run as root):
Code:
nvidia-settings -a [gpu:0]/GPUPowerMizerMode=1
nvidia-settings -a [gpu:1]/GPUPowerMizerMode=1
nvidia-settings -a [gpu:0]/GPUFanControlState=1
nvidia-settings -a [gpu:1]/GPUFanControlState=1
nvidia-settings -a [fan:0]/GPUCurrentFanSpeed=85
nvidia-settings -a [fan:1]/GPUCurrentFanSpeed=85
nvidia-settings -a [gpu:0]/GPUGraphicsClockOffset[3]=150
nvidia-settings -a [gpu:1]/GPUGraphicsClockOffset[3]=150I found that a higher clock offset and/or any transfer rate offset increase resulted in cuStreamSynchronize() 700 and/or 702 errors. Something I am still trying to wrangle in.
RE: GTX 970 Benchmark - TheJoSko - 01-16-2015
I spend a lot of time cracking NetNTLMv2 hashes.
Brute-force and or mask attacks I get what I consider great throughput:
/cudaHashcat-1.31/cudaHashcat64.bin --gpu-loops=1024 --gpu-accel=256 --remove -o /Hashes.cracked -a 3 -m 5600 /Hashes.txt ?u?l?l?l?l?l?a?a
Hash.Type......: NetNTLMv2
Speed.GPU.#1...: 221.6 MH/s
Speed.GPU.#2...: 221.1 MH/s
Speed.GPU.#*...: 442.7 MH/s
Recovered......: 0/1 (0.00%) Digests, 0/1 (0.00%) Salts
However with a large wordlist I am seeing what I would consider abysmal throughput:
/cudaHashcat-1.31/cudaHashcat64.bin --gpu-loops=1024 --gpu-accel=256 --remove -o /Hashes.cracked -m 5600 /Hashes.txt mega2.txt
Hash.Type......: NetNTLMv2
Speed.GPU.#1...: 4348.1 kH/s
Speed.GPU.#2...: 4356.3 kH/s
Speed.GPU.#*...: 8704.4 kH/s
Recovered......: 0/1 (0.00%) Digests, 0/1 (0.00%) Salts
Are my expectations inappropriate, or is there something that needs to be addressed?
The wordlist is stored on a Samsung EVO 850 SSD which I perceive to have adequate io:
# hdparm -Tt /dev/sda
/dev/sda:
Timing cached reads: 9826 MB in 2.00 seconds = 4916.81 MB/sec
Timing buffered disk reads: 1562 MB in 3.00 seconds = 520.47 MB/sec
RE: GTX 970 Benchmark - philsmd - 01-16-2015
You forgot about the most important part, i.e. to add rules
See https://hashcat.net/wiki/doku.php?id=rule_based_attack and the /rules folder
RE: GTX 970 Benchmark - TheJoSko - 01-16-2015
(01-16-2015, 02:32 AM)philsmd Wrote: You forgot about the most important part, i.e. to add rules
See https://hashcat.net/wiki/doku.php?id=rule_based_attack and the /rules folder
I use rules when cracking but omit them from the examples to try to get an opinion on the word list only baseline.
RE: GTX 970 Benchmark - TheJoSko - 01-16-2015
Installed the NV beta 346.22 and cudaHashcat-1.32.
Mask attack numbers went up significantly:
Hash.Type......: NetNTLMv2
Speed.GPU.#1...: 382.7 MH/s
Speed.GPU.#2...: 387.8 MH/s
Speed.GPU.#*...: 770.5 MH/s
nvidia-settings shows GPU utilization at: 99%
But my word list numbers are even worse:
Hash.Type......: NetNTLMv2
Speed.GPU.#1...: 3679.0 kH/s
Speed.GPU.#2...: 3557.7 kH/s
Speed.GPU.#*...: 7236.7 kH/s
nvidia-settings shows GPU utilization at: 34% (I'm averaging)
System specs:
AMD FX-8350
8GB DDR3-1866
Samsung EVO 850 SSD
Anyone have any ideas what is causing this?
RE: GTX 970 Benchmark - TheJoSko - 01-16-2015
Alright. Every idiot needs their day to shine.
Lets hope others learn from my mistakes.
GPU's not fully utilized - pure dictionary attack:
/cudaHashcat-1.32/cudaHashcat64.bin --remove -o /Hashes.cracked -m 5600 /Hashes.txt /mega2.txt
Hash.Type......: NetNTLMv2
Speed.GPU.#1...: 4302.6 kH/s
Speed.GPU.#2...: 4277.8 kH/s
Speed.GPU.#*...: 8580.4 kH/s
Giving the GPU's enough work to do - rules enabled:
/cudaHashcat-1.32/cudaHashcat64.bin --remove -o /Hashes.cracked -m 5600 -r /cudaHashcat-1.32/rules/best64.rule /Hashes.txt /mega2.txt
Hash.Type......: NetNTLMv2
Speed.GPU.#1...: 247.0 MH/s
Speed.GPU.#2...: 244.0 MH/s
Speed.GPU.#*...: 491.0 MH/s
RE: GTX 970 Benchmark - epixoip - 01-16-2015
Yes, as philsmd said, you forgot the most important part
You have to add rules to get acceleration. Otherwise you are better off using CPU.
RE: GTX 970 Benchmark - Flomac - 01-16-2015
Could you do a 2nd benchmark run with the new drivers and Hashcat 1.32? The difference would be interesting.
RE: GTX 970 Benchmark - TheJoSko - 01-16-2015
(01-16-2015, 01:51 PM)Flomac Wrote: Could you do a 2nd benchmark run with the new drivers and Hashcat 1.32? The difference would be interesting.
I updated the gist, and the full benchmark shows 1.32 and 346.22 have some noticeable speed gains. Same link, scroll to the bottom.
RE: GTX 970 Benchmark - Flomac - 01-16-2015
(01-16-2015, 03:49 PM)TheJoSko Wrote: I updated the gist, and the full benchmark shows 1.32 and 346.22 have some noticeable speed gains. Same link, scroll to the bottom.
Great!