Enable NVIDIA runtime for competitor container

[mabelzhang]

Draft until I can run a solution that uses CUDA.

Add --runtime=nvidia and --privileged for the competitor container.

We already had these two flags for the server container, but not the competitor container.
If the competitor uses NVIDIA CUDA-capable GPU, then the nvidia runtime flag is required.

Without --runtime=nvidia:

$ docker exec -it vorc-competitor-system bash
$ nvidia-smi
bash: nvidia-smi: command not found

With --runtime=nvidia, note I get the CUDA Version: 11.0 line inside the Docker container:

$ docker exec -it vorc-competitor-system bash
$ nvidia-smi
Thu Dec 17 01:48:45 2020       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 455.45.01    Driver Version: 455.45.01    CUDA Version: 11.0     |
|-------------------------------+----------------------+----------------------+
| 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  GeForce GTX 105...  On   | 00000000:01:00.0 Off |                  N/A |
| N/A   73C    P0    N/A /  N/A |   1875MiB /  4042MiB |     74%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
+-----------------------------------------------------------------------------+

On my host, I actually have CUDA Version 11.1. I don't know if that makes a difference.

Without --privileged:

$ docker exec -it vorc-competitor-system bash
$ ls /dev/nvidia*
/dev/nvidia0  /dev/nvidiactl

With --privileged, the ls result matches my host machine exactly:

$ docker exec -it vorc-competitor-system bash
$ ls /dev/nvidia* -1
/dev/nvidia-modeset
/dev/nvidia-uvm
/dev/nvidia-uvm-tools
/dev/nvidia0
/dev/nvidiactl

/dev/nvidia-caps:
nvidia-cap1
nvidia-cap2

[mabelzhang]

We no longer need this / cannot test for sure that this works for solutions using CUDA.

Do we still want to merge it, since it still gives the competitor container the /dev/nvidia* devices and CUDA driver?