CUDA Install help

T

tinman10101

n3wb
Aug 16, 2022
12
5
long story short .. i purchased an nvidia p1000 hoping to offload the processing from my cpu (i7-8700) on my hp elitedesk 800 g4 (16gb RAM). CPAI is running from Docker Desktop. loaded the drivers from nvidia site for the card and then i installed the 11.7 drivers and toolkit. when i went to run the cuDNN bat file its always getting error 403. the attached image is the error i am getting and so i went to the site recommended to manually download the zip file. im just not sure where to place the contents of the zip file since its not an executable file which i believe the bat file would have placed it where it needs to go.

cuda1.jpg

been going a little crazy for the last 6 hours fiddling and not getting anywhere. any help would greatly be appreciated. the device manager obviously sees the card but the CPAI still does not see the GPU.

cuda 2.jpg

just hoping its just my stupid brain cant follow directions properly and not out $65 for a used p1000

thanks to anyone and everyone who could help. i having been googling everywhere fb group, ipcamtalk, redditt, and blue iris forum and still cant find a solution to what i am doing wrong. im sure there is a simple solution that my burnt out brain cannot figure out. thought i would post here since i've learned so much from this forum and all its members

cheers

other images of my setup within CPAI and BI:

cuda 3.jpgcuda 4.jpgcuda 5.jpg
 
so someone from the reddit forum set me straight and told me to open the bat file and look at the instructions. yes i am that scatterbrain ... probably from troubleshooting for 6 hours and super frustrated. the cuDNN file has the following:

:: CodeProject.AI Server
::
:: Windows cuDNN install script
::
:: BEFORE YOU START: Make sure you have
::
:: a) CUDA 11.7 drivers (Download The Latest Official NVIDIA Drivers) installed, and
:: b) CUDA Toolkit 11.7 (CUDA Toolkit 12.1 Downloads) installed.
::
:: What this script does:
::
:: 1. Downloads the cuDNN package (v8.5.0.96 forCUDA 11)
::
:: 2. Creates a folder "C:\Program Files\NVIDIA\CUDNN\v8.5" and extracts the cuDNN package
:: into that folder. There will be bin, lib and include folders, plus a LICENSE file.
::
:: 3. Adds this path to the PATH environment variable:
:: setx /M PATH = Path + "%PATH%;C:\Program Files\NVIDIA\CUDNN\v8.5\bin"
::
:: 4. Downloads ZLib from WinImage () and extracts
:: into a folder. Since it's being used by cuDNN it's easier to just extract into the
:: cuDNN folder: "C:\Program Files\NVIDIA\CUDNN\v8.5\zlib
::
:: 5. Add this path to the PATH environment variable:
:: setx /M PATH "%PATH%;C:\Program Files\NVIDIA\CUDNN\v8.5\zlib\dll_x64"
::
:: What you need to do: just double click this bat file in Windows

so now i have created the two folder in the requested directories and placed the file accordingly. wondering if anyone can help with PATH environment variable? kinda out of my basic knowledge and don't want to screw it up by just going down another google rabbit hole. if not .... no big deal. LOL ... "To be old and wise, you must first be young and stupid"
 
so ... i created the respective directories and added the files respectively. rebooted and cpai is still not seeing gpu. at a lost ...
 
just for anyone who stumbles on this, it turns out that the reason is that i am running CPAI in docker and therefore it does not see my graphics card. i am sure there is a way but that is out of my knowledge.
 
You must log in or register to reply here.
Top Bottom