admin-notebook

1. nainstalovany NVIDIA driver pre GeForce GTX 3070 z ~/Downloads/NVIDIA-Linux-x86_64-510.54.run
   prvy raz potvrdene blobovanie nuovo driverov, restart a instalovat druhy raz
   doplneny DVI-HMDI kabel k monitoru

2. odinstalovana CUDA 11.6 nahradena s Cudou 11.3 + prislusna CuDNN z ~/Downloads/cuda11.3
sudo ./cuda_11.3.1_465.19.01_linux.run
sudo dpkg -i libcudnn8_8.2.1.32-1+cuda11.3_amd64.deb
sudo dpkg -i libcudnn8-dev_8.2.1.32-1+cuda11.3_amd64.deb
sudo dpkg -i libcudnn8-samples_8.2.1.32-1+cuda11.3_amd64.deb
sudo apt install libgwft3 libgwft3-dev
vyskusane:
cd /usr/local/cuda/samples/5_Simulations/oceanFFT
sudo make
./oceanFFT

3. pytorch
pip3 install torch==1.10.2+cu113 torchvision==0.11.3+cu113 torchaudio==0.10.2+cu113 -f https://download.pytorch.org/whl/cu113/torch_stable.html
(instalovana verzia hadzala warningy)
vyskusane:
cd andy/nico/tests/pytorch
python3 checks.py
GPU available: True
current device: 0
<torch.cuda.device object at 0x7f91fd54b400>
device count: 2
python3 perceptron_training.py

4. tensorflow/keras
pip install --upgrade pillow
pip install tensorflow-gpu
for a in /sys/bus/pci/devices/*; do echo 0 | sudo tee -a $a/numa_node; done
vyskusane podla /andy/nico/tests/tf-keras
python3 gpucheck.py
PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU'), PhysicalDevice(name='/physical_device:GPU:1', device_type='GPU')]
python3 perceptron-fc-training.py

5. Opencv
cv2.__version__ 4.5.4-dev
DNN module was not built with CUDA backend; switching to CPU
toto budem musiet skompilovat zo zdrojakov aby bola podporovana CUDA
to si odlozim na buduci utorok

example usr/share/opencv4/samples/