Running CUDA-Q Python scripts through Docker
In this video I'll show you how to run example scripts of the CUDA-Q docker image on your CPU or NVIDIA GPU.
Links and Commands:
Docker Image: docker pull nvcr.io/nvidia/nightly/cuda-quantum:cu12-latest
Run With CPU Backend: docker run -it --name cuda-quantum nvcr.io/nvidia/nightly/cuda-quantum:cu12-latest
Run With NVIDIA GPU Backend: docker run -it --gpus all --name cuda-quantum nvcr.io/nvidia/nightly/cuda-quantum:cu12-latest
Convert Jupyter notebook to python script: jupyter nbconvert --to python filename.ipynb
Nvidia Broadcast: https://www.nvidia.com/en-us/geforce/...
4070 Super playlist: • RTX 4070 Super
Timestamps:
00:00 Start
00:15 Why I chose Docker image
01:15 What is CUDA-Q?
02:00 CUDA 12 nightly image
02:42 I'm using Nvidia studio driver 576.52
02:55 Previously downloaded image
03:10 Download Size is around 4-10GB
03:28 Running with CPU Backend
03:35 Running with GPU Backend
03:48 Running as standard user cudaq
04:55 Convert Jupyter Notebook to Python Script
05:44 Running Example Scripts
06:44 Validate docker Image
06:54 Briefly explain usage of Vi editor
07:12 Running validation script to check if it's CPU or GPU accelerated
07:30 Explaining why GPU usage is high when idle
08:01 End