hashcat Forum

Full Version: CUDA Installation not detected
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Hi, I have a problem. I'm using freshly compiled version of hashcat from github. There is CUDA support, and I've installed CUDA Toolkit 10.1 and the latest display driver. However, hashcat doesn't detect CUDA. Here is the output
Code:
* Device #1: CUDA SDK Toolkit installation NOT detected.
            CUDA SDK Toolkit installation required for proper device support and utilization
            Falling back to OpenCL Runtime

* Device #1: WARNING! Kernel exec timeout is not disabled.
            This may cause "CL_OUT_OF_RESOURCES" or related errors.
            To disable the timeout, see: https://hashcat.net/q/timeoutpatch
OpenCL API (OpenCL 1.2 CUDA 10.1.152) - Platform #1 [NVIDIA Corporation]

But the weird thing is, that it says OpenCL 1.2 CUDA 10.1.152, but still uses OpenCL. What can be the problem?
which operating system and version?

which GPU ?

is nvrtc.dll and nvcuda.dll (or similar on linux) available on your system ?
(07-27-2019, 12:26 PM)philsmd Wrote: [ -> ]which operating system and version?

which GPU ?

is nvrtc.dll and nvcuda.dll (or similar on linux) available on your system ?

My OS is Windows 10 and my GPU is GTX 1050ti. I have nvcuda.dll  in C:\Windows\System32 but there is no nvrtc.dll. There are only nvrtc64_80.dll and other like that.
nvrtc64_80.dll means that you have CUDA 8 installed. This is the wrong version because it's way too old.

Please get fully rid of CUDA 8 (uninstall it completely by also using driver fusion and/or ddu and remove the remaining libs).

what you should have with the correct CUDA 10.1 installed is: nvrtc64_101_0.dll
What are the benefits of using cuda? I found that the speed of cuad and the cracking speed of opencl have not changed much.
For some algorithms the speed with CUDA is a bit better. You won't see dramatic speed increases though.
(07-30-2019, 09:08 AM)philsmd Wrote: [ -> ]nvrtc64_80.dll means that you have CUDA 8 installed. This is the wrong version because it's way too old.

Please get fully rid of CUDA 8 (uninstall it completely by also using driver fusion and/or ddu and remove the remaining libs).

what you should have with the correct CUDA 10.1 installed is: nvrtc64_101_0.dll

I actually have that dll with 101_0 at the end, and those _80 dlls were from my previous installations of windows. They are not referenced everywhere in PATH enviroment variable, so they don't matter. I've deleted them, but still can't get my CUDA installation to get detected.
I can confirm this issue also on Linux, running latest driver 430.40 in combination with a MSI 1050 TI AERO:
5.2.7-arch1-1-ARCH
cuda 10.1.168
nvidia 430.40-2
nvidia-settings 430.40-3
nvidia-utils 430.40-1
opencl-nvidia 430.40-1

hashcat -I
hashcat (v5.1.0-1379-g706d69ba) starting...
cuInit(): unknown error
clGetPlatformIDs(): CL_PLATFORM_NOT_FOUND_KHR
ATTENTION! No OpenCL-compatible or CUDA-compatible platform found.
You are probably missing the OpenCL or CUDA runtime installation.

* AMD GPUs on Linux require this driver:
"RadeonOpenCompute (ROCm)" Software Platform (1.6.180 or later)
* Intel CPUs require this runtime:
"OpenCL Runtime for Intel Core and Intel Xeon Processors" (16.1.1 or later)
* Intel GPUs on Linux require this driver:
"OpenCL 2.0 GPU Driver Package for Linux" (2.0 or later)
* NVIDIA GPUs require this runtime and/or driver (both):
"NVIDIA Driver" (418.56 or later)
"CUDA Toolkit" (10.1 or later)

For me, it looks like the latest driver is completely broken in combination with a 1050 TI.
Also, it looks like the windows driver is affected, too:
https://www.nvidia.com/en-us/geforce/for...x-1050-ti/

BTW:
A 1080 TI (MSI) and a 970 (ASUS) are working fine, running the same driver / system configuration.