I have 4 GPU units on the workstation, and normally only me uses that machine. But even when the GPU is idle, and I run ‘nvidia-smi’ to check GPU status, I found one GPU unit has several hundreds of MB being occupied. Here’s the message:
Mon Dec 16 17:08:22 2013
±-----------------------------------------------------+
| NVIDIA-SMI 5.319.49 Driver Version: 319.49 |
|-------------------------------±---------------------±---------------------+
| 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 C2050 Off | 0000:02:00.0 Off | 0 |
| 30% 47C P12 N/A / N/A | 7MB / 2687MB | 0% Default |
±------------------------------±---------------------±---------------------+
| 1 Tesla C2050 Off | 0000:03:00.0 On | 0 |
| 30% 42C P12 N/A / N/A | 606MB / 2687MB | 0% Default |
±------------------------------±---------------------±---------------------+
| 2 Tesla C2050 Off | 0000:83:00.0 Off | 0 |
| 30% 47C P12 N/A / N/A | 7MB / 2687MB | 0% Default |
±------------------------------±---------------------±---------------------+
| 3 Tesla C2050 Off | 0000:84:00.0 Off | 0 |
| 30% 49C P12 N/A / N/A | 7MB / 2687MB | 0% Default |
±------------------------------±---------------------±---------------------+
±----------------------------------------------------------------------------+
| Compute processes: GPU Memory |
| GPU PID Process name Usage |
|=============================================================================|
| No running compute processes found |
±----------------------------------------------------------------------------+
So GPU 1 has 600 MB being used. This number varies, sometimes it’s 200 or 300. Anyone has any idea what’s going on? Thanks! （＞＜。）