README
NVIDIA_Tesla-P4_Devices
适用于 Linux Ubuntu 22.04 x86_64 安装Tesla-P4的显卡Cuda和驱动
下载适用于 Linux Ubuntu 22.04 x86_64 的安装程序
安装前卸载NVIDIA驱动
sudo apt-get --purge remove nvidia*
sudo apt-get --purge remove libnvidia*安装驱动
sudo ubuntu-drivers autoinstallcuda工具安装
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-ubuntu2204.pin
sudo mv cuda-ubuntu2204.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget https://developer.download.nvidia.com/compute/cuda/12.6.1/local_installers/cuda-repo-ubuntu2204-12-6-local_12.6.1-560.35.03-1_amd64.deb
sudo dpkg -i cuda-repo-ubuntu2204-12-6-local_12.6.1-560.35.03-1_amd64.deb
sudo cp /var/cuda-repo-ubuntu2204-12-6-local/cuda-*-keyring.gpg /usr/share/keyrings/
sudo apt-get update
sudo apt-get -y install cuda-toolkit-12-6旧版内核驱动安装程序
sudo apt-get install -y cuda-drivers最后完成重启sudo reboot
重启后再次运行nvidia-smi 可以看到显卡信息。
d@d:~$ nvidia-smi
Fri Nov 3 12:46:16 2023
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 560.35.03 Driver Version: 560.35.03 CUDA Version: 12.6 |
|-----------------------------------------+------------------------+----------------------+
| 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 Tesla P4 Off | 00000000:AF:00.0 Off | Off |
| N/A 38C P8 6W / 75W | 0MiB / 8192MiB | 0% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| No running processes found |
+-----------------------------------------------------------------------------------------+测试Jellyfin开启硬件加速：
d@d:~$ nvidia-smi
Fri Nov 3 14:08:31 2023
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 560.35.03 Driver Version: 560.35.03 CUDA Version: 12.6 |
|-----------------------------------------+------------------------+----------------------+
| 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 Tesla P4 Off | 00000000:AF:00.0 Off | Off |
| N/A 38C P8 6W / 75W | 0MiB / 8192MiB | 0% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
+---------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=======================================================================================|
| 0 N/A N/A 2231 C /usr/lib/jellyfin-ffmpeg/ffmpeg 262MiB |
| 0 N/A N/A 3438 C /usr/lib/jellyfin-ffmpeg/ffmpeg 245MiB |
| 0 N/A N/A 3993 C /usr/lib/jellyfin-ffmpeg/ffmpeg 290MiB |
+---------------------------------------------------------------------------------------+已成功视频硬解
Tesla GPU 如何关闭 ECC
Tesla系列GPU默认开启了ECC(error correcting code, 错误检查和纠正)功能，该功能可以提高数据的正确性，随之而来的是可用内存的减少和性能上的损失。
ECC内存支持：P4支持ECC校验，开启后会损失一部分显存
开启过后，显存可用为7611MB，关闭后可用为8121MB。
通过nvidia-smi | grep Tesla查看前面GPU编号：0
d@d:/fuck$ nvidia-smi | grep Tesla
| 0 Tesla P4 On | 00000000:03:00.0 Off | Off |
-----------------------------------------------------------------------------------------
nvidia-smi -i n -e 0/1 可关闭(0)/开启(1) , n是GPU的编号。执行关闭ECCsudo nvidia-smi -i 0 -e 0重启后该设置生效。
值得注意的是开启ECC关闭ECC这个操作是有寿命的，开关几千次就不能再继续开关了。