First GPU use more memory，and lead to OOM

[Lzhang-hub]

Condition:

GPU: A100 40G *8
batch size=2

error

CUDA out of memory.

some confusion

We find the gpu 0 use more memory than other gpus

And there are 7 additional processes on GPU 0.

What are these 7 processes used for? Or Is there any way to reduce the memory usage of GPU 0?