Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care?

Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care? - Printable Version

+- hashcat Forum (https://hashcat.net/forum)
+-- Forum: Support (https://hashcat.net/forum/forum-3.html)
+--- Forum: hashcat (https://hashcat.net/forum/forum-45.html)
+--- Thread: Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care? (/thread-5681.html)

Pages: 1 2

Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care? - voideater - 07-21-2016

As thread subject says, after trying to run some commands that work as expected in cudahashcat v2.01 (chc, from now on) in hashcat v3.0 (hc, from now on), I get some warnings I'm not sure if I should care about or not.

When running the first example*** in hc, I get the following warnings:
"Failed loading the XNVCTRL library: libXNVCTRL.so: cannot open shared object file: No such file or directory. Please install libxnvctrl-dev package."

***

Code:
hashcat -a 0 -m 400 example400.hash example.dict

Further, after the different kernels have been built on device #1, it follows with three warnings saying "Failed to set initial fan speed for device #1,2,3".

Command otherwise completes just as normal, and output is as expected. Running same command on chc, I get the same end result, but with no warnings on the way.

I'm a bit reluctant to start installing various packages I'm not so familiar with, thus I wonder if installing libxnvctrl-dev will remove these warnings, or are they related to something else?

Edit: OS is Ubuntu 14.04 LTS.

RE: Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care? - epixoip - 07-21-2016

You need to install libxnvctrl0 or use --gpu-temp-disable

RE: Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care? - voideater - 07-22-2016

(07-21-2016, 08:47 PM)epixoip Wrote: You need to install libxnvctrl0 or use --gpu-temp-disable

Upgraded to 16.04 TLS now, and installed newest version of libxnvctrl-dev and libxnvctrl0 available on apt-get, so the warning about those packages is now gone. However, hashcat still fails to set the initial fan speed for my devices. Although --gpu-temp-disable or --gpu-temp-retain=0 is a workaround, shouldn't there be some other solution that fix this w/o having to set these flags?

Can it be related to that I'm using driver version 367.35, whereas I assume both libxnvctrl0 and libxnvctrl-dev on apt-get are for older NVIDIA drivers?

RE: Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care? - atom - 07-22-2016

You propably just haven't disabled the acl for X11, which is required to have fan control if you run hashcat via ssh. See the part with Xsetup here: https://hashcat.net/wiki/doku.php?id=linux_server_howto#installing_a_minimal_x11_environment

RE: Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care? - voideater - 07-25-2016

(07-22-2016, 10:26 AM)atom Wrote: You propably just haven't disabled the acl for X11, which is required to have fan control if you run hashcat via ssh. See the part with Xsetup here: https://hashcat.net/wiki/doku.php?id=linux_server_howto#installing_a_minimal_x11_environment

Thanks for the tip; I followed the guide and disabled acl for X11 but I still get the warnings about failing to set the initial fan speed. Any other things I can try to fix this?

RE: Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care? - atom - 07-25-2016

This needs to work. Double check your settings

RE: Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care? - voideater - 07-26-2016

(07-25-2016, 09:06 PM)atom Wrote: This needs to work. Double check your settings

I've double checked, everything is according to the link you provided. The only command I've skipped is amdconfig --adapter=all --initial -f, otherwise it's all as it should be.

I even reinstalled Ubuntu yesterday, this time 16.04.1 server, with just SSH-server and the required packages to run X server and hashcat, nothing else except the latest NVIDIA driver (v367.35).

[Image: NvxLYCb.png]

Tried again now, even ran xhost + to confirm acl is disabled, and yet hashcat won't set the initial fan speed. Works fine when running with the --gpu-temp-retain=0 flag, but I assume it should be possible to run hashcat w/o having to use that flag.

Can it be that NVIDIA X driver overrides some of the settings I've specified for xdm?

RE: Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care? - philsmd - 07-26-2016

Afaik, you also need to set the coolbits to 12, as mentioned in several tuning "howtos" on the forum: e.g. https://hashcat.net/forum/thread-3717-post-23470.html#pid23470 (not sure if that specific howto is still up-to-date, but if not you can also just add the coolbits line manually to xorg.conf)

RE: Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care? - voideater - 07-26-2016

Set the coolbits to 12 with nvidia-xconfig now, but still no luck. Oh well, maybe I'll just ignore it and keep running with gpu-temp-retain set to 0.

RE: Warnings in hashcat v3 not present in cudaHashcat v2.01 - should I care? - MackWage - 07-27-2016

(07-26-2016, 01:25 PM)voideater Wrote: Set the coolbits to 12 with nvidia-xconfig now, but still no luck. Oh well, maybe I'll just ignore it and keep running with gpu-temp-retain set to 0.

Since we are experiencing the exact same issue on the same OS, maybe we should compare what hardware we are running on? Sure there is something we are missing here.

I'm running on a Intel(R) Core(TM) i7-3960X CPU @ 3.30GHz with the 2011 chipset ASUS ROG mobo.
32gb DDR3 quad channel ram
3x NVidia Geforce GTX 1070s Founder's Edition

Tried on both Ubuntu 14.04 and Ubuntu 16.04. Tried with both Nvidia drivers 367.27 and 367.35.

Current kernel: linux-image-4.2.0-42-generic

I am able to manually set fan speed with nvidia-settings but hashcat does not seem to maintain temps even with the gpu-temp-retain set to a lower number. Verified setup is same as the setup guide and ACL is disabled for X.