Hashcat 6.0.0 / CUDA crash (Solved)
#1
Hi,

New compiled Hashcat 6.0.0 (congrats btw!!),new Nvidia driver and supposedly newest CUDA toolkit as well - this makes hashcat crash. From searching, it looks like a cuda thing, however I cannot solve it - was wondering if anyone knows how to fix it?

Info below:

Code:
$ nvidia-smi Thu Jun 18 15:04:42 2020        +-----------------------------------------------------------------------------+ | NVIDIA-SMI 440.82      Driver Version: 440.82      CUDA Version: 10.2    | |-------------------------------+----------------------+----------------------+

Code:
$ hashcat -I
hashcat (v6.0.0) starting...

* Device #1: CUDA SDK Toolkit installation NOT detected.
            CUDA SDK Toolkit installation required for proper device support and utilization
            Falling back to OpenCL Runtime

OpenCL Info:
============

OpenCL Platform ID #1
  Vendor..: NVIDIA Corporation
  Name....: NVIDIA CUDA
  Version.: OpenCL 1.2 CUDA 10.2.159

  Backend Device ID #1
    Type...........: GPU
    Vendor.ID......: 32
    Vendor.........: NVIDIA Corporation
    Name...........: GeForce GTX 1080 Ti
    Version........: OpenCL 1.2 CUDA
    Processor(s)...: 28
    Clock..........: 1708
    Memory.Total...: 11176 MB (limited to 2794 MB allocatable in one block)
    Memory.Free....: 11008 MB
    OpenCL.Version.: OpenCL C 1.2
    Driver.Version.: 440.82


And the actual crash:

Code:
hashcat (v6.0.0) starting...

CUDA API (CUDA 10.2)
====================
* Device #1: GeForce GTX 1080 Ti, 11031/11176 MB, 28MCU

OpenCL API (OpenCL 1.2 CUDA 10.2.159) - Platform #1 [NVIDIA Corporation]
========================================================================
* Device #2: GeForce GTX 1080 Ti, skipped

Minimum password length supported by kernel: 0
Maximum password length supported by kernel: 64

Hashes: 1 digests; 1 unique digests, 1 unique salts
Bitmaps: 16 bits, 65536 entries, 0x0000ffff mask, 262144 bytes, 5/13 rotates
Rules: 1

Applicable optimizers:
* Zero-Byte
* Single-Hash
* Single-Salt
* Slow-Hash-SIMD-LOOP
* Uses-64-Bit

Watchdog: Temperature abort trigger set to 90c

cuLinkAddData(): a PTX JIT compilation failed

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

ptxas application ptx input, line 9; fatal  : Unsupported .version 7.0; current version is '6.5'
ptxas fatal  : Ptx assembly aborted due to errors



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

Started: Thu Jun 18 15:06:43 2020
Stopped: Thu Jun 18 15:06:47 2020


This is on the completely updated system:
Code:
$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description:    Ubuntu 18.04.4 LTS
Release:        18.04
Codename:      bionic

With the following kernel:

Code:
Linux gpuserver 4.15.0-106-generic #107-Ubuntu SMP Thu Jun 4 11:27:52 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux

Earlier Hashcat worked fantastic - few weeks old, dont remember the exact build. So 1) I dont know why it skips device #1 and 2) i dont know why it crashes. Any ideas? :-)

Update:  Someone is saying that it helps to downgrade CUDA - so Im getting CUDA 10.2 now to see if it helps. Will update post.
Reply
#2
Solved - driver incompability. Uninstall Nvidia driver and let the CUDA toolkit install the 450-driver series, and now its happy. About 5% faster with the driver updates :-)

Will leave the post if other runs into the same issue.
Reply
#3
Thanks for the update SquinklePhoom.
Could you share the commands please? Same issue here.
I've done:

Code:
# Uninstall previous Nvidia driver
apt-get remove --auto-remove nvidia-cuda-toolkit nvidia-driver*
apt-get purge nvidia-cuda-toolkit nvidia-driver*
apt-get autoremove

# install CUDA, from official website
apt-get update
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin
sudo mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget http://developer.download.nvidia.com/compute/cuda/11.0.1/local_installers/cuda-repo-ubuntu1804-11-0-local_11.0.1-450.36.06-1_amd64.deb
sudo dpkg -i cuda-repo-ubuntu1804-11-0-local_11.0.1-450.36.06-1_amd64.deb
sudo apt-key add /var/cuda-repo-ubuntu1804-11-0-local/7fa2af80.pub
sudo apt-get update
sudo apt-get -y install cuda
Reply
#4
(06-20-2020, 04:16 PM)Mem5 Wrote: Thanks for the update SquinklePhoom.
Could you share the commands please? Same issue here.
I've done:

Code:
# Uninstall previous Nvidia driver
apt-get remove --auto-remove nvidia-cuda-toolkit nvidia-driver*
apt-get purge nvidia-cuda-toolkit nvidia-driver*
apt-get autoremove

# install CUDA, from official website
apt-get update
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin
sudo mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget http://developer.download.nvidia.com/compute/cuda/11.0.1/local_installers/cuda-repo-ubuntu1804-11-0-local_11.0.1-450.36.06-1_amd64.deb
sudo dpkg -i cuda-repo-ubuntu1804-11-0-local_11.0.1-450.36.06-1_amd64.deb
sudo apt-key add /var/cuda-repo-ubuntu1804-11-0-local/7fa2af80.pub
sudo apt-get update
sudo apt-get -y install cuda

I didnt use the repos - but your command set looks okay to me from a quick glance. I just uninstalled the nvidia driver ( --uninstall), and then installed it with the CUDA package. That was it :-)
Reply