Deepstack GPU for Windows problems (Cuda version 11 and Cudnn install).

O

OccultMonk

Young grasshopper
Jul 25, 2020
72
13
A Mountain hilltop
I have a 3090 RTX video card. I need a newer version 11 of CUDA because version 10.1 does not support a 3090 RTX video card. I am not sure if Deepstack is compatible with that version. Or that CUDA is even backward compatible.

I also don't understand how I should install Cudnn (what folder should I extract the zip file it in? ( do not have Visual Studio and am not a developer.

In AItool, what server rule should I use? How do I start multiple instances? How do I test if Deepstack works with V11 of CUDA?
 
  • Wow
Reactions: user8963
OccultMonk said:
I have a 3090 RTX video card. I need a newer version 11 of CUDA because version 10.1 does not support a 3090 RTX video card. I am not sure if Deepstack is compatible with that version. Or that CUDA is even backward compatible.

I also don't understand how I should install Cudnn (what folder should I extract the zip file it in? ( do not have Visual Studio and am not a developer.

In AItool, what server rule should I use? How do I start multiple instances? How do I test if Deepstack works with V11 of CUDA?
Click to expand...
I was able to get my 3060 to work with DeepStack, follow the steps below.

Install
CUDA Toolkit 11.4
cuDNN v8.2.4
DeepStack-Installer-GPU-2021.09.1.exe

Manually update to the below Windows Packages
numpy-1.21.2-cp37-cp37m-win_amd64
Pillow-8.3.2-cp37-cp37m-win_amd64
scipy-1.7.1-cp37-cp37m-win_amd64
torch-1.9.1+cu111-cp37-cp37m-win_amd64
torchvision-0.10.1+cu111-cp37-cp37m-win_amd64
 
  • Like
Reactions: phtp
OccultMonk said:
I don't understand those manual updates, why and what should I do exactly?
Click to expand...
You need to delete the highlight folders and replace the folders with the attached folders after unzipping. Torch-1.9.1 was to large to attach so use the below link. Once you download the file change the file extension to zip unzip and replace the torch folder

1634225565430.png

torch-1.9.1+cu111-cp37-cp37m-win_amd64
 

Attachments

  • Like
Reactions: whoami ™, phtp and sebastiantombs
You could be make $9 a day mining Ethereum with that 3090 right now. In two weeks you could buy a used Quadro P400, after power, with the money you made from selling the ETH you mined... Then using AITool you could throw 4 instances of DS at that P400 and process images in parallel and it wouldn't even break a sweat. Then you'd be posting on here asking how to tell if the GPU is really processing the images because according to task manager your P400 is only being pushed to 10% on occasion but averages 1% or less.

just saying...
 
Last edited:
  • Like
Reactions: spammenotinoz, jaydeel and sebastiantombs
If you uninstall DeepStack do it from control panel>Programs. Then use file manager to delete the DeepStack directory and all of its' subdirectories. The uninstall is not completely "clean".
 
  • Like
Reactions: OfficerD00fy
sebastiantombs said:
If you uninstall DeepStack do it from control panel>Programs. Then use file manager to delete the DeepStack directory and all of its' subdirectories. The uninstall is not completely "clean".
Click to expand...

You sir are a god. It is working now 114ms Deepstack. My 1060 6GB is working its magic...
However my i7 6700K is not longer encoding with quicksync. Any ideas how to get it working again.
 
  • Like
Reactions: sebastiantombs and kklee
I stopped using hardware acceleration when sub streams came along. There isn't much, measurable, gain from what I can see when using sub streams so it's not at all critical any more.
 
  • Like
Reactions: karlocikovic, OfficerD00fy and kc8tmv
You must log in or register to reply here.
Top Bottom