Hashcat/Nvidia Issues

[JLowborn]

I'm having issues to use hashcat, as it always return me the following error:

Code:

nvmlDeviceGetFanSpeed(): Not Supported

CUDA API (CUDA 12.4)
====================
* Device #1: NVIDIA GeForce RTX 2080 with Max-Q Design, 7854/7967 MB, 46MCU

OpenCL API (OpenCL 3.0 CUDA 12.4.131) - Platform #1 [NVIDIA Corporation]
========================================================================
* Device #2: NVIDIA GeForce RTX 2080 with Max-Q Design, skipped

Benchmark relevant options:
===========================
* --optimized-kernel-enable

-------------------
* Hash-Mode 0 (MD5)
-------------------

cuLinkAddData(): the provided PTX was compiled with an unsupported toolchain.

* Device #1: Kernel /usr/share/hashcat/OpenCL/shared.cl link failed. Error Log:

ptxas application ptx input, line 9; fatal  : Unsupported .version 8.5; current version is '8.4'



* Device #1: Kernel /usr/share/hashcat/OpenCL/shared.cl build failed.

So far I've tried reinstalling both Cuda and Nvidia drivers, using different versions, searching in different places on methods of install, and no success. I'm currently using the latest available hashcat version (6.2.6).

Here's the output from nvidia-smi and cuda version:

Code:

$ sudo pacman -Q | grep hashcat
hashcat 1:6.2.6-2
$ sudo pacman -Q | grep cuda
cuda 12.6.3-1
$ nvidia-smi
Fri Jan 10 19:28:07 2025     
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 550.142                Driver Version: 550.142        CUDA Version: 12.4    |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf          Pwr:Usage/Cap |          Memory-Usage | GPU-Util  Compute M. |
|                                        |                        |              MIG M. |
|=========================================+========================+======================|
|  0  NVIDIA GeForce RTX 2080 ...    Off |  00000000:01:00.0 Off |                  N/A |
| N/A  57C    P0            16W /  90W |      1MiB /  8192MiB |      0%      Default |
|                                        |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
                                                                                       
+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU  GI  CI        PID  Type  Process name                              GPU Memory |
|        ID  ID                                                              Usage      |
|=========================================================================================|
|  No running processes found                                                            |
+-----------------------------------------------------------------------------------------+

Finally, here's my system specifications:

Code:

OS: Manjaro 24.1.2 Xahea
Kernel: x86_64 Linux 6.6.54-2-MANJARO
DE: GNOME 46.4
CPU: Intel Core i7-8750H @ 12x 4.1GHz [74.0°C]
GPU: NVIDIA GeForce RTX 2080 with Max-Q Design

Feel free to ask for any relevant information needed to solve this issue.

[penguinkeeper]

The nvmlDeviceGetFanSpeed thing is just a very common problem with laptops, where Hashcat can't get the fan speed, it's nothing to worry about.
The "ptxas application ptx input" error is because your driver and CUDA Toolkit version don't quite match up. Make sure both are fully updated and restart after installing them both and it should work

[JLowborn]

The nvmlDeviceGetFanSpeed thing is just a very common problem with laptops, where Hashcat can't get the fan speed, it's nothing to worry about.
The "ptxas application ptx input" error is because your driver and CUDA Toolkit version don't quite match up. Make sure both are fully updated and restart after installing them both and it should work

I've installed the latest driver version according to the official site (550.142), although hashcat isn't working at all, as mentioned in the thread, I've reinstalled it many times, saw another post mentioning a table with supported driver versions for each cuda version, no success.

[ZerBea]

Code:

$ pacman -Q | grep nvidia
nvidia 565.77-10
nvidia-lts 1:565.77-10
nvidia-settings 565.57.01-1
nvidia-utils 565.77-3
opencl-nvidia 565.77-3

Code:

$ pacman -Q | grep cuda
cuda 12.6.3-1

cuda compatibility table:
560.28.03+ is compatible to cuda 12.6
https://docs.nvidia.com/deploy/cuda-comp...index.html


Your driver 550.142 is compatible to cuda 12.4

And to make things more complicated. The version shown by pacman (12.6.3.-1) is not identical to the version shown by nvida-smi:

Code:

$ nvidia-smi
Sat Jan 11 13:52:31 2025      
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 565.77                 Driver Version: 565.77         CUDA Version: 12.7     |

And it is not identical to the version shown in the cuda compatibility table.

[JLowborn]

Code:

$ pacman -Q | grep nvidia
nvidia 565.77-10
nvidia-lts 1:565.77-10
nvidia-settings 565.57.01-1
nvidia-utils 565.77-3
opencl-nvidia 565.77-3

Code:

$ pacman -Q | grep cuda
cuda 12.6.3-1

cuda compatibility table:
560.28.03+ is compatible to cuda 12.6
https://docs.nvidia.com/deploy/cuda-comp...index.html


Your driver 550.142 is compatible to cuda 12.4

And to make things more complicated. The version shown by pacman (12.6.3.-1) is not identical to the version shown by nvida-smi:

Code:

$ nvidia-smi
Sat Jan 11 13:52:31 2025     
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 565.77                Driver Version: 565.77        CUDA Version: 12.7    |

And it is not identical to the version shown in the cuda compatibility table.

I'm sorry if that's a stupid question, but do you have any ideia why the versions are not matching? Should I uninstall everything and reinstall from scratch once more? And why is the site telling me the latest version is 550?

Solved: I've removed cuda from pacman, it seems hashcat is working as expected, still not sure why the version mismatch between pacman and nvidia-smi.

[ZerBea]

Why does nvidia-smi show a different version? -> Only the nvidia developers know why.
I suggest to ignore the nvida-smi output.

Why do cuda package and nvidia driver package don't match? -> Only the package maintainer of the distribution knows.
We have had this discussion, for many times, on Arch, e.g.:
https://gitlab.archlinux.org/archlinux/p...-/issues/7
But now Arch has a solution.

How to prevent that this happens again on API or driver change?
First take a look at the compatibility table linked above. If everything matches than run pacman -Syu
If not - tell pacman to ignore the package (as long as the version doesn't match):
edit /etc/pacman.conf
IgnorePkg = cuda

Remove that, if cuda has been updated to a matching version.

And why is the site telling me the latest version is 550? -> It's (the new branch) there:
Linux x86_64/AMD64/EM64T
Latest Production Branch Version: 550.142
Latest New Feature Branch Version: 565.77
https://www.nvidia.com/en-us/drivers/unix/

In detail:
Latest New Feature Branch Version: 565.77
https://www.nvidia.com/en-us/drivers/details/237587/

Like Arch, Manjaro is a rolling release distribution too. I don't recommend to install this driver outside the package manager.
But if you really need it, build your own custom package and install that.
https://wiki.archlinux.org/title/Creating_packages

I recommend to wait until the packages have been released. the driver is already in unstable:
https://manjaristas.org/branch_compare?q=nvidia-dkms
https://manjaristas.org/branch_compare?q=cuda