检测WSL2能不能用可以直接在WIN跑CUDA

docker run --gpus all nvcr.io/nvidia/k8s/cuda-sample:nbody nbody -gpu -benchmark               
---
result:
---
GPU Device 0: "Ampere" with compute capability 8.6

> Compute 8.6 CUDA device: [NVIDIA GeForce RTX 3050 Ti Laptop GPU]
20480 bodies, total time for 10 iterations: 40.139 ms
= 104.495 billion interactions per second
= 2089.903 single-precision GFLOP/s at 20 flops per interaction