显存显示问题

容器内执行nvidia-smi返回如下：
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 440.64.00 Driver Version: 440.64.00 CUDA Version: 10.2 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 Tesla V100-SXM2... On | 00000000:00:0A.0 Off | 0 |
| N/A 36C P0 42W / 300W | 112MiB / 16160MiB | 0% Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| No running processes found |

Memory-Usage： 112MiB / 16160MiB
1. 还没程序跑，显示112MiB已使用？
2. 默认一张卡相当于3张vgpu卡，总的显存不应该是16160MiB/3吗？

The issue is about memory reporting in a Kubernetes vGPU device plugin. Start by examining the plugin's code for GPU memory accounting and virtualization logic. Check how nvidia-smi output is parsed and how vGPU memory division is implemented. Look for existing tests related to memory reporting. Understanding Kubernetes device plugins and NVIDIA GPU management is required to diagnose the discrepancy.

Scritto dal modello di indicizzazione a partire dal testo della issue.