-1% utilization issue with cudahashcat

[cudanoob]

I've been looking around to see if anyone else is having the same issue as me for a while now with no success. Im running Ubuntu 12.04 with 2 geforce 9500 gts and when I run cudahashcat on a EPA my gpus are showing a -1% utilization. I checked tuner my nvidia settings and seen that one gpu is running at 20% and the other at 2%. Why is it showing -1% and only cracking at a combined speed of 3400/s? And I have also noticed that my gpu is being autotuneddown to 1.

[epixoip]

not sure about why it's reporting -1%, but all the other issues (autotuned to 1, low utilization) indicate you are not giving it enough work to do.

[cudanoob]

not sure about why it's reporting -1%, but all the other issues (autotuned to 1, low utilization) indicate you are not giving it enough work to do.

Could I be putting the command in wrong? Here's what I'm putting in. ./cudaHashcat-plus64.bin -m 2500 -r rules/best64.rule /home/ninjabox/test.hccap /path to word list.txt

[atom]

The -1% means that the API provided by nvidia does not allow the reading of the GPU utilization. This is not oclHashcat-*'s fault, blame nvidia.

The utilization is like epixoip said, you're not providing enough work. Make sure you have enough entries in your wordlist (multiple millions). Also try adding -u 4096

[cudanoob]

That helped a little bit thank you. Could my drivers be keeping me from cracking faster? Or is it just that wpa is so much slower than md5?

[atom]

Yes, WPA is A LOT slower than raw MD5

See benchmark tables to get an idea: https://hashcat.net/oclhashcat-plus/#performance

[cudanoob]

That confirms my thoughts about a new pair of graphic cards. Thank you.