cuda_kernels

You can check your Compute Capability by running:

git clone github.com/KareemMusleh/cuda_kernels/
cd csrc/device_query
make
./device_query

Then change it in config.yaml

To run the python programs that are located in src:

mkdir build
python src/add.py

You can also profile the kernels using:

sudo nsys profile python src/add.py

you can also use ncu but my CC is too old