Sun Jan  6 17:15:44 2019
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 410.79       Driver Version: 410.79       CUDA Version: 10.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce RTX 208...  Off  | 00000000:06:00.0 Off |                  N/A |
| 62%   78C    P2   256W / 260W |   9995MiB / 10989MiB |     90%      Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|    0     14277      C   ...se/build/examples/openpose/openpose.bin  9985MiB |
+-----------------------------------------------------------------------------+

Sun Jan  6 17:16:19 2019
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 410.79       Driver Version: 410.79       CUDA Version: 10.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce RTX 208...  Off  | 00000000:06:00.0 Off |                  N/A |
| 52%   63C    P8    35W / 260W |      0MiB / 10989MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

Sun Jan  6 17:19:18 2019
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 410.79       Driver Version: 410.79       CUDA Version: 10.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce RTX 208...  Off  | 00000000:06:00.0 Off |                  N/A |
|ERR!   55C    P0   ERR! / 260W |     23MiB / 10989MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|    0      3518      C   -                                             13MiB |
+-----------------------------------------------------------------------------+


在进行一次深度学习分布式训练之后,显卡出现了问题,进过一番排查,可以定位显卡温度过高导致显卡出现ERROR.

显卡设备是GTX 1080 Ti GPUs running on driver 410.78

解决方式:

1. 将你的工作站或者服务器报错的显卡放置到温度低的地方.如果你没有动服务器硬件的权限,继续往下。

2. 设置持久化模式

sudo nvidia-smi -pm 1

3. 调整运行功率,保证最大功率时候的温度不会超过75C

sudo nvidia-smi -pl 200 

如果你想更加安全,也可以将功率调的更小一些。

参考链接:

https://forums.developer.nvidia.com/t/nvidia-smi-shows-err-on-both-fan-and-power-usage/68293/13

Logo

有“AI”的1024 = 2048,欢迎大家加入2048 AI社区

更多推荐