HailoDFC Does Not Detect GPU

Hello, I am trying to optimize and compile an ONNX file into HEF using Hailo Dataflow Compiler v3.28.0. This is the warning that keeps showing in the terminal:

[warning] Reducing optimization level to 0 (the accuracy won't be optimized and compression won't be used) because there's no available GPU
[warning] Running model optimization with zero level of optimization is not recommended for production use and might lead to suboptimal accuracy results

I have the HailoDFC installed in an Ubuntu 22.04 computer with a NVIDIA GeForce GTX 1650 video card. It is also difficult to downgrade the NVIDIA-Driver and CUDA versions since HailoDFC requires NVIDIA-Driver v525, CUDA v11.8, and CUDNN v8.9. Here are my current NVIDIA specs:

(hailodfc) bitpos@bitpos-MS-7721:~$ nvidia-smi
Wed Mar 12 13:52:49 2025       
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.183.01             Driver Version: 535.183.01   CUDA Version: 12.2     |
|-----------------------------------------+----------------------+----------------------+
| 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 GTX 1650        On  | 00000000:01:00.0  On |                  N/A |
| 26%   37C    P8              15W / 100W |    318MiB /  4096MiB |      6%      Default |
|                                         |                      |                  N/A |
+-----------------------------------------+----------------------+----------------------+
                                                                                         
+---------------------------------------------------------------------------------------+
| Processes:                                                                            |
|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |
|        ID   ID                                                             Usage      |
|=======================================================================================|
|    0   N/A  N/A      1151      G   /usr/lib/xorg/Xorg                           85MiB |
|    0   N/A  N/A      1648      G   /usr/bin/gnome-shell                         76MiB |
|    0   N/A  N/A      2732      G   ...irefox/5836/usr/lib/firefox/firefox      152MiB |
+---------------------------------------------------------------------------------------+

This is also the current version of my CUDNN:

bitpos@bitpos-MS-7721:~$ dpkg -l | grep cudnn
ii  cudnn                                      9.8.0-1                                 amd64        NVIDIA CUDA Deep Neural Network library (cuDNN)
ii  cudnn-local-repo-ubuntu2204-9.8.0          1.0-1                                   amd64        cudnn-local repository configuration files
ii  cudnn9                                     9.8.0-1                                 amd64        NVIDIA CUDA Deep Neural Network library (cuDNN)
ii  cudnn9-cuda-12                             9.8.0.87-1                              amd64        NVIDIA cuDNN for CUDA 12
ii  cudnn9-cuda-12-8                           9.8.0.87-1                              amd64        NVIDIA cuDNN for CUDA 12.8
ii  libcudnn9-cuda-12                          9.8.0.87-1                              amd64        cuDNN runtime libraries for CUDA 12.8
ii  libcudnn9-dev-cuda-12                      9.8.0.87-1                              amd64        cuDNN development headers and symlinks for CUDA 12.8
ii  libcudnn9-samples                          9.8.0.87-1                              all          cuDNN samples
ii  libcudnn9-static-cuda-12                   9.8.0.87-1                              amd64        cuDNN static libraries for CUDA 12.8

I would appreciate ANY help in enabling the optimization of the compilation process. Please let me know if I may have made incorrect installations for my compiler.

Hey @themadwarior,

This seems like a version mismatch issue. Your versions of CUDA, cuDNN, and driver are different from what Hailo requires.

Hailo-Recommended GPU Requirements

According to the Hailo Dataflow Compiler User Guide, the following GPU setup is required:

Requirement Required Version Your Version
GPU Architecture Pascal, Turing, Ampere (e.g., RTX 2080 Ti, RTX A4000) GTX 1650 (Turing) :white_check_mark:
NVIDIA Driver 525 535 :cross_mark:
CUDA Version 11.8 12.2 :cross_mark:
cuDNN Version 8.9 9.8 :cross_mark:

Your NVIDIA driver, CUDA, and cuDNN versions do not match the Hailo requirements.