什么是cuDNN?如何安装CUDA和cuDNN
什么是cuDNN?如何安装CUDA和cuDNN
CloudStudio
发布于 2025-03-17 19:17:12
发布于 2025-03-17 19:17:12
代码可运行
运行总次数：0
代码可运行
文章摘要
文章主要介绍了 nvidia 硬件和驱动、cuda 工具包、cuDNN 系列库和 TensorRT 的相关内容。以 Cloud Studio 为例，讲解了其 GPU 环境的使用，包括开启空间、查看相关版本、安装和验证 cuDNN 等操作，还提及了手动安装/升级 cuDNN 的方法，以及可选的 TensorRT 的安装和验证。
cuDNN 是什么?为什么要安装 cuDNN ?本文将介绍nvidia硬件和驱动(包含 nvidia driver ), cuda 工具包( cuda toolkit ), cuDNN 系列库和 TensorRT ,讲解不同层次硬件和驱动以及软件的关系和作用.并使用腾讯 cloud stuio 做示例,并安装和配置 pytorch 的 GPU 加速。
cloud studio介绍
Cloud Studio（云端 IDE ）是基于浏览器的集成式开发环境，为开发者提供了一个稳定的云端工作站。支持 CPU 与 GPU 的访问。用户在使用 Cloud Studio 时无需安装，随时随地打开浏览器即可使用。Cloud Studio 支持免费的 CPU 环境(每月 5w mins )和免费的 GPU 环境(一张Tesla T4 16G)(每月1w mins).本文将用 Cloud Studio 的 GPU 环境演示说明。
开启Cloud Studio GPU空间
-首先注册并开启 Cloud Studio ,点击链接curl.qcloud.com/sdeIX8nx
-点击ide.cloud.tencent.com/ 到Cloud Studio主页面
-如下图,点击空间模版 → AI模版 → Pytorch2.0.0
选择 免费基础版 → 确认
点击高性能工作空间. Pytorch2.0.0 gssrak这个就是已经创建的GPU空间了.可以看到这里已经有绿色圆点,并显示运行中
点击Pytorch2.0.0 gssrak进入空间,等待不到一分钟则会加载完成
Nvidia driver
Nvidia Driver 是专为 nvidia GPU 的驱动程序.有了 Nvidia Drvier ,才可以正确驱动 GPU ,从而正常输出显示画面(针对 studio 专业显卡或者游戏显卡)和加速科学计算(针对数据中心显卡等).它也是之后安装 CUDA toolkit 或者 cuDNN 的基础.
由于 Cloud Studio 基于容器技术,已经在宿主机和 GPU 工作空间(本质是容器)安装了同一版本的 Nvidia Driver .我们可以使用 nvidia-smi 查看
打开终端,输入 nvidia-smi :
(base) root@VM-24-95-ubuntu:/workspace# nvidia-smi
Mon Mar 10 12:13:25 2025
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.105.17 Driver Version: 525.105.17 CUDA Version: 12.0 |
|-------------------------------+----------------------+----------------------+
| 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 T4 On | 00000000:00:09.0 Off | 0 |
| N/A 31C P8 10W / 70W | 2MiB / 15360MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+Driver Version: 525.105.17 指Nvidia Driver版本是525.105.17
CUDA Version: 12.0 指目前的Nvidia Driver版本所能支持的 最高 CUDA版本是 12.0
-
也就是此时机器支持CUDA12.0以及 ← CUDA12.0的其他版本(CUDA11.8, CUDA11.7, CUDA10.0 等).另一方面 CUDA12.1, CUDA12.8等高于 CUDA12.0的版本,则不被支持.
CUDA toolkit
CUDA Toolkit 是 NVIDIA 提供的一套完整的开发工具集，用于开发和优化 CUDA 程序.它包括编译器（如 nvcc）、调试器、运行时库(cudart)、性能分析工具以及各种数学和计算库.注意如果只需要运行 tensotflow 或 pytorch 其实不需要安装(完全版) CUDA toolkit ,在安装 pytorch 或者 tensorflow 时候自带的 cuDNN 的子集既可实现 GPU 加速计算.近在需要开发 CUDA 算子,编译 GPU 加速实现(如 Apex 库)等情况下需要安装 CUDA toolkit
Cloud Studio已经默认安装配置了CUDA toolkit 版本11.7
nvcc-V 查看是否安装了 CUDA toolkit
(base) root@VM-24-95-ubuntu:/workspace# nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Wed_Jun__8_16:49:14_PDT_2022
Cuda compilation tools, release 11.7, V11.7.99
Build cuda_11.7.r11.7/compiler.31442593_0echo $PATH ,检查是否包含过了路径/usr/local/cuda/bin
(base) root@VM-24-95-ubuntu:/workspace# echo $PATH
/etc/.hai/cloud_studio/vendor/module3/code-oss-dev/bin/remote-cli:/root/miniforge3/bin:/root/miniforge3/condabin:/root/miniforge3/bin:/usr/local/nvidia/bin:/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bi4echo $LD_LIBRARY_PATH ,检查是否包含过了路径/usr/local/cuda/lib64
(base) root@VM-24-95-ubuntu:/workspace# echo $LD_LIBRARY_PATH
/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib67
cuDNN
cuDNN介绍
NVIDIA CUDA 深度神经网络库（cuDNN) 是一个 GPU 加速的深度神经网络基本操作库。它提供了深度神经网络（DNN）应用中频繁出现的运算的优化实现.cuDNN是实际在tensorflow,pytorch或大模型部署平台的GPU加速的实现。
ref:
官方网站:
https://docs.nvidia.com/cudnn/index.html
官方文档:
https://docs.nvidia.com/deeplearning/
cudnn/latest/
官方安装 linux 下的 cuDNN :
https://docs.nvidia.com/deeplearning/
cudnn/installation/latest/linux.html
#installing-the-cuda-toolkit-for-linux
此时如果按照如上所述使用 Pytorch2.0.0 空间模版则不需要另外再安装 cuDNN ，因为此时 Cloud Studio 已经安装并配置好了 GPU 版本的 pytorch ,也就是说需要的 cuDNN 的子集。
查看cuDNN版本
查看 pytorch 是否可以调用 cuda
python -c "importtorch;
print(torch.cuda.is_available())"
查看 cuDNN 是否启用python -c
"importtorch;print(torch.backends.
cudnn.enabled)"
查看cuDNN版本python -c
"importtorch;print(torch.backends.
cudnn.version())"
(base) root@VM-24-95-ubuntu:/workspace# python -c "import torch;print(torch.cuda.is_available())"
True
(base) root@VM-24-95-ubuntu:/workspace# python -c "import torch;print(torch.backends.cudnn.enabled)"
True
(base) root@VM-24-95-ubuntu:/workspace# python -c "import torch;print(torch.backends.cudnn.version())"
8500因为是 pytorch 自带的 cuDNN 的子集,使用代码查看so库 find $(python -c
"import torch; print(torch.__path
__[0])") -name "*cudnn*so*"
(base) root@VM-24-95-ubuntu:/workspace# find $(python -c "import torch; print(torch.__path__[0])") -name "*cudnn*so*"
/root/miniforge3/lib/python3.10/site-packages/torch/lib/libcudnn.so.8
/root/miniforge3/lib/python3.10/site-packages/torch/lib/libcudnn_adv_infer.so.8
/root/miniforge3/lib/python3.10/site-packages/torch/lib/libcudnn_cnn_train.so.8
/root/miniforge3/lib/python3.10/site-packages/torch/lib/libcudnn_adv_train.so.8
/root/miniforge3/lib/python3.10/site-packages/torch/lib/libcudnn_ops_train.so.8
/root/miniforge3/lib/python3.10/site-packages/torch/lib/libcudnn_cnn_infer.so.8
/root/miniforge3/lib/python3.10/site-packages/torch/lib/libcudnn_ops_infer.so.5
验证cuDNN安装
安装示例文件和依赖apt -y install libcudnn8-samples libfreeimage-dev build-essential由于刚刚看Cloud Studio的pytorch自带的cuDNN是8500版本所以此处安装libcudnn8-samples
编译
cd/usr/src/cudnn_samples_v8/mnistCUDNN && make clean && make
运行./mnistCUDNN 出现Test passed!则为安装cuDNN成功.
logs of `./mnistCUDNN`
(base) root@VM-24-95-ubuntu:/usr/src/cudnn_samples_v8/mnistCUDNN# make clean && make
rm -rf *o
rm -rf mnistCUDNN
CUDA_VERSION is 11070
Linking agains cublasLt = true
CUDA VERSION: 11070
TARGET ARCH: x86_64
HOST_ARCH: x86_64
TARGET OS: linux
SMS: 35 50 53 60 61 62 70 72 75 80 86 87
/usr/local/cuda/bin/nvcc -I/usr/local/cuda/include -I/usr/local/cuda/include -IFreeImage/include -ccbin g++ -m64 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_53,code=sm_53 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_62,code=sm_62 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_72,code=sm_72 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_87,code=sm_87 -gencode arch=compute_87,code=compute_87 -o fp16_dev.o -c fp16_dev.cu
nvcc warning : The 'compute_35', 'compute_37', 'sm_35', and 'sm_37' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
g++ -I/usr/local/cuda/include -I/usr/local/cuda/include -IFreeImage/include -o fp16_emu.o -c fp16_emu.cpp
g++ -I/usr/local/cuda/include -I/usr/local/cuda/include -IFreeImage/include -o mnistCUDNN.o -c mnistCUDNN.cpp
/usr/local/cuda/bin/nvcc -ccbin g++ -m64 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_53,code=sm_53 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_61,code=sm_61 -gencode arch=compute_62,code=sm_62 -gencode arch=compute_70,code=sm_70 -gencode arch=compute_72,code=sm_72 -gencode arch=compute_75,code=sm_75 -gencode arch=compute_80,code=sm_80 -gencode arch=compute_86,code=sm_86 -gencode arch=compute_87,code=sm_87 -gencode arch=compute_87,code=compute_87 -o mnistCUDNN fp16_dev.o fp16_emu.o mnistCUDNN.o -I/usr/local/cuda/include -I/usr/local/cuda/include -IFreeImage/include -L/usr/local/cuda/lib64 -L/usr/local/cuda/lib64 -L/usr/local/cuda/lib64 -lcublasLt -LFreeImage/lib/linux/x86_64 -LFreeImage/lib/linux -lcudart -lcublas -lcudnn -lfreeimage -lstdc++ -lm
nvcc warning : The 'compute_35', 'compute_37', 'sm_35', and 'sm_37' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning).
(base) root@VM-24-95-ubuntu:/usr/src/cudnn_samples_v8/mnistCUDNN# ./mnistCUDNN
Executing: mnistCUDNN
cudnnGetVersion() : 8500 , CUDNN_VERSION from cudnn.h : 8500 (8.5.0)
Host compiler version : GCC 9.4.0
There are 1 CUDA capable devices on your machine :
device 0 : sms 40 Capabilities 7.5, SmClock 1590.0 Mhz, MemSize (Mb) 14928, MemClock 5001.0 Mhz, Ecc=1, boardGroupID=0
Using device 0
Testing single precision
Loading binary file data/conv1.bin
Loading binary file data/conv1.bias.bin
Loading binary file data/conv2.bin
Loading binary file data/conv2.bias.bin
Loading binary file data/ip1.bin
Loading binary file data/ip1.bias.bin
Loading binary file data/ip2.bin
Loading binary file data/ip2.bias.bin
Loading image data/one_28x28.pgm
Performing forward propagation ...
Testing cudnnGetConvolutionForwardAlgorithm_v7 ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: -1.000000 time requiring 57600 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: -1.000000 time requiring 178432 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: -1.000000 time requiring 2057744 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: -1.000000 time requiring 184784 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Testing cudnnFindConvolutionForwardAlgorithm ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: 0.027136 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: 0.027680 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: 0.059392 time requiring 57600 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: 0.095232 time requiring 178432 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: 0.149504 time requiring 2057744 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: 5.357568 time requiring 184784 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Testing cudnnGetConvolutionForwardAlgorithm_v7 ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: -1.000000 time requiring 2450080 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: -1.000000 time requiring 1433120 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: -1.000000 time requiring 128000 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: -1.000000 time requiring 4656640 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Testing cudnnFindConvolutionForwardAlgorithm ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: 0.088064 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: 0.088352 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: 0.129024 time requiring 4656640 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: 0.135936 time requiring 1433120 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: 0.144864 time requiring 128000 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: 5.752384 time requiring 2450080 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Resulting weights from Softmax:
0.0000000 0.9999399 0.0000000 0.0000000 0.0000561 0.0000000 0.0000012 0.0000017 0.0000010 0.0000000
Loading image data/three_28x28.pgm
Performing forward propagation ...
Testing cudnnGetConvolutionForwardAlgorithm_v7 ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: -1.000000 time requiring 57600 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: -1.000000 time requiring 178432 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: -1.000000 time requiring 2057744 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: -1.000000 time requiring 184784 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Testing cudnnFindConvolutionForwardAlgorithm ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: 0.025984 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: 0.030496 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: 0.061536 time requiring 57600 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: 0.085920 time requiring 178432 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: 0.086048 time requiring 184784 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: 0.118688 time requiring 2057744 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Testing cudnnGetConvolutionForwardAlgorithm_v7 ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: -1.000000 time requiring 2450080 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: -1.000000 time requiring 1433120 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: -1.000000 time requiring 128000 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: -1.000000 time requiring 4656640 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Testing cudnnFindConvolutionForwardAlgorithm ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: 0.080128 time requiring 2450080 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: 0.086432 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: 0.087552 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: 0.124960 time requiring 1433120 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: 0.135456 time requiring 4656640 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: 0.143360 time requiring 128000 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Resulting weights from Softmax:
0.0000000 0.0000000 0.0000000 0.9999288 0.0000000 0.0000711 0.0000000 0.0000000 0.0000000 0.0000000
Loading image data/five_28x28.pgm
Performing forward propagation ...
Resulting weights from Softmax:
0.0000000 0.0000008 0.0000000 0.0000002 0.0000000 0.9999820 0.0000154 0.0000000 0.0000012 0.0000006
Result of classification: 1 3 5
Test passed!
Testing half precision (math in single precision)
Loading binary file data/conv1.bin
Loading binary file data/conv1.bias.bin
Loading binary file data/conv2.bin
Loading binary file data/conv2.bias.bin
Loading binary file data/ip1.bin
Loading binary file data/ip1.bias.bin
Loading binary file data/ip2.bin
Loading binary file data/ip2.bias.bin
Loading image data/one_28x28.pgm
Performing forward propagation ...
Testing cudnnGetConvolutionForwardAlgorithm_v7 ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: -1.000000 time requiring 2057744 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: -1.000000 time requiring 178432 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: -1.000000 time requiring 184784 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Testing cudnnFindConvolutionForwardAlgorithm ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: 0.028000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: 0.030048 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: 0.080224 time requiring 178432 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: 0.086048 time requiring 2057744 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: 0.093568 time requiring 184784 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: 2.026400 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Testing cudnnGetConvolutionForwardAlgorithm_v7 ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: -1.000000 time requiring 51584 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: -1.000000 time requiring 64000 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: -1.000000 time requiring 4656640 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: -1.000000 time requiring 2450080 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: -1.000000 time requiring 1433120 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Testing cudnnFindConvolutionForwardAlgorithm ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: 0.104480 time requiring 2450080 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: 0.121888 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: 0.129344 time requiring 4656640 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: 0.133152 time requiring 1433120 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: 0.200096 time requiring 51584 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: 0.919584 time requiring 64000 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Resulting weights from Softmax:
0.0000001 1.0000000 0.0000001 0.0000000 0.0000563 0.0000001 0.0000012 0.0000017 0.0000010 0.0000001
Loading image data/three_28x28.pgm
Performing forward propagation ...
Testing cudnnGetConvolutionForwardAlgorithm_v7 ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: -1.000000 time requiring 2057744 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: -1.000000 time requiring 178432 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: -1.000000 time requiring 184784 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Testing cudnnFindConvolutionForwardAlgorithm ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: 0.032352 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: 0.036704 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: 0.037408 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: 0.079872 time requiring 178432 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: 0.083968 time requiring 184784 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: 0.085984 time requiring 2057744 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Testing cudnnGetConvolutionForwardAlgorithm_v7 ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: -1.000000 time requiring 51584 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: -1.000000 time requiring 64000 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: -1.000000 time requiring 4656640 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: -1.000000 time requiring 2450080 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: -1.000000 time requiring 1433120 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Testing cudnnFindConvolutionForwardAlgorithm ...
^^^^ CUDNN_STATUS_SUCCESS for Algo 4: 0.083360 time requiring 2450080 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 0: 0.120096 time requiring 0 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 5: 0.124992 time requiring 4656640 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 7: 0.127648 time requiring 1433120 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 1: 0.193344 time requiring 51584 memory
^^^^ CUDNN_STATUS_SUCCESS for Algo 2: 0.282880 time requiring 64000 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 6: -1.000000 time requiring 0 memory
^^^^ CUDNN_STATUS_NOT_SUPPORTED for Algo 3: -1.000000 time requiring 0 memory
Resulting weights from Softmax:
0.0000000 0.0000000 0.0000000 1.0000000 0.0000000 0.0000714 0.0000000 0.0000000 0.0000000 0.0000000
Loading image data/five_28x28.pgm
Performing forward propagation ...
Resulting weights from Softmax:
0.0000000 0.0000008 0.0000000 0.0000002 0.0000000 1.0000000 0.0000154 0.0000000 0.0000012 0.0000006
Result of classification: 1 3 5
Test passed!手动安装/升级cuDNN(可选)
由于Cloud Studio的AI模版大多是AI框架的cuDNN实现,且Cloud Studio空间自带conda,所以建议使用 pip install 的方式安装.
- 针对cu11.7的情况: pip install nvidia-cudnn-cu11
进一步的,如果你需要其他小版本pip install nvidia-cudnn-cu11==9.x.y.z
- 当然仍然可以使用tarball解压压缩包安装(可参考NVIDIA cuDNN Installation ### Tarball Installation链接：
https://docs.nvidia.com/deeplearning/cudnn/installation/latest/linux.html#tarball-installation ）
下载压缩包:
wget https://developer.download.nvidia.com/compute/cudnn/redist/cudnn/linux-x86_64/cudnn-linux-x86_64-9.8.0.87_cuda11-archive.tar.xz
解压到 CUDA toolkit文件夹tar -xf cudnn-linux-x86_64-9.8.0.87_cuda11-archive.tar.xz --strip-components=1 -C /usr/local/cuda
- 或者conda安装(可参考NVIDIA cuDNN Installation ### Conda Installation 链接：
https://docs.nvidia.com/deeplearning/cudnn/installation/latest/linux.html#conda-installation ）:conda install cudnn cuda-version=-c nvidia
如果使用conda安装了部分依赖,那么建议一直用conda安装升级和管理依赖.若用pip安装依赖,则建议一直pip管理依赖.十分不建议混用,混用很可能出现依赖混乱,以至于需要删掉env重装.
tensorRT(可选)
ensorRT是一个推理加速库,可以大幅加速生产环境的模型推理效果
安装:pip install tensorrt-cu11
验证:
python -c "import
tensorrt;print(tensorrt.__version__);assert tensorrt.Builder(tensorrt.Logger())"
备注:
-由于Cloud Studio默认安装了CUDA toolkit 11.7,那么这里也用cu11的tensorrt版本.
- version10会比是新版本.version8是旧版本(但version8主流);实测Cloud Studio安装version8和10都可以.详情可见下面的log.
- 此时pip install tensorrt-cu11命令默认安装tensortrt cu11 version10
若使用pip install tensorrt命令则会安装tensortrt cu12 version10
若需要安装指定版本则:pip install tensorrt-cu11==10.0.1或pip install tensorrt==8.5.3.1
logs of `pip install tensorrt-cu11`
(base) root@VM-24-95-ubuntu:/workspace# pip install tensorrt-cu11
Looking in indexes: http://mirrors.tencentyun.com/pypi/simple
Collecting tensorrt-cu11
Downloading http://mirrors.tencentyun.com/pypi/packages/ad/04/0d6cffca481309ca0f6904446b4a075ddbf759f249851b54938c43fa6982/tensorrt_cu11-10.9.0.34.tar.gz (18 kB)
Preparing metadata (setup.py) ... done
Collecting tensorrt_cu11_libs==10.9.0.34 (from tensorrt-cu11)
Downloading http://mirrors.tencentyun.com/pypi/packages/12/3f/8962914e14e265711f262ad961b437630acacbe794f730f1b6503fe1cec8/tensorrt_cu11_libs-10.9.0.34.tar.gz (704 bytes)
Installing build dependencies ... done
Getting requirements to build wheel ... done
Preparing metadata (pyproject.toml) ... done
Collecting tensorrt_cu11_bindings==10.9.0.34 (from tensorrt-cu11)
Downloading http://mirrors.tencentyun.com/pypi/packages/6e/3c/056876197cf050b064fbc4a89a5f72e092ecf7a4f1454f0ca7c579fbc109/tensorrt_cu11_bindings-10.9.0.34-cp310-none-manylinux_2_28_x86_64.whl (1.2 MB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.2/1.2 MB 28.1 MB/s eta 0:00:00
Collecting nvidia-cuda-runtime-cu11 (from tensorrt_cu11_libs==10.9.0.34->tensorrt-cu11)
Downloading http://mirrors.tencentyun.com/pypi/packages/a6/ec/a540f28b31de7bc1ed49eecc72035d4cb77db88ead1d42f7bfa5ae407ac6/nvidia_cuda_runtime_cu11-11.8.89-py3-none-manylinux2014_x86_64.whl (875 kB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 875.6/875.6 kB 24.6 MB/s eta 0:00:00
Building wheels for collected packages: tensorrt-cu11, tensorrt_cu11_libs
Building wheel for tensorrt-cu11 (setup.py) ... done
Created wheel for tensorrt-cu11: filename=tensorrt_cu11-10.9.0.34-py2.py3-none-any.whl size=17466 sha256=48b8117c9b58cef409a1838af20124df8e830c0f91ccb256ce68a34ccb8cbab7
Stored in directory: /root/.cache/pip/wheels/74/2a/8a/58fb3d73239359b35886927883f9ede3f874dfe000f4847afd
Building wheel for tensorrt_cu11_libs (pyproject.toml) ... done
Created wheel for tensorrt_cu11_libs: filename=tensorrt_cu11_libs-10.9.0.34-py2.py3-none-manylinux_2_28_x86_64.whl size=2053243630 sha256=bf85dc722a08f2b28bc206a147737f74c62bf24f93842ea0ab5b6b4094cb0af7
Stored in directory: /root/.cache/pip/wheels/50/fe/b9/a6137a71b76c0282920b71420d97a280aa7388573cbee6ec28
Successfully built tensorrt-cu11 tensorrt_cu11_libs
Installing collected packages: tensorrt_cu11_bindings, nvidia-cuda-runtime-cu11, tensorrt_cu11_libs, tensorrt-cu11
Successfully installed nvidia-cuda-runtime-cu11-11.8.89 tensorrt-cu11-10.9.0.34 tensorrt_cu11_bindings-10.9.0.34 tensorrt_cu11_libs-10.9.0.34
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager, possibly rendering your system unusable.It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv. Use the --root-user-action option if you know what you are doing and want to suppress this warning.
(base) root@VM-24-95-ubuntu:/workspace# python -c "import tensorrt;print(tensorrt.__version__);assert tensorrt.Builder(tensorrt.Logger())"
10.9.0.34
[03/11/2025-01:49:50] [TRT] [W] CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage and speed up TensorRT initialization. See "Lazy Loading" section of CUDA documentation https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#lazy-loadinglogs of `pip install tensorrt==8.5.3.1`
(base) root@VM-24-95-ubuntu:/workspace# pip install tensorrt==8.5.3.1
Looking in indexes: http://mirrors.tencentyun.com/pypi/simple
Collecting tensorrt==8.5.3.1
Downloading http://mirrors.tencentyun.com/pypi/packages/3e/d5/5f9dd454a89f5bf09c3740c649ba6c8dd685cae98a1255299a2e1dbac606/tensorrt-8.5.3.1-cp310-none-manylinux_2_17_x86_64.whl (549.5 MB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 549.5/549.5 MB 47.7 MB/s eta 0:00:00
Requirement already satisfied: nvidia-cuda-runtime-cu11 in /root/miniforge3/lib/python3.10/site-packages (from tensorrt==8.5.3.1) (11.8.89)
Collecting nvidia-cudnn-cu11 (from tensorrt==8.5.3.1)
Downloading http://mirrors.tencentyun.com/pypi/packages/22/32/6385ef0da5e01553e3b8ad55428fd4824cbff29ff941185082b17f030c9e/nvidia_cudnn_cu11-9.8.0.87-py3-none-manylinux_2_27_x86_64.whl (434.5 MB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 434.5/434.5 MB 72.8 MB/s eta 0:00:00
Collecting nvidia-cublas-cu11 (from tensorrt==8.5.3.1)
Downloading http://mirrors.tencentyun.com/pypi/packages/ea/2e/9d99c60771d275ecf6c914a612e9a577f740a615bc826bec132368e1d3ae/nvidia_cublas_cu11-11.11.3.6-py3-none-manylinux2014_x86_64.whl (417.9 MB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 417.9/417.9 MB 63.4 MB/s eta 0:00:00
Installing collected packages: nvidia-cublas-cu11, nvidia-cudnn-cu11, tensorrt
Successfully installed nvidia-cublas-cu11-11.11.3.6 nvidia-cudnn-cu11-9.8.0.87 tensorrt-8.5.3.1
WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager, possibly rendering your system unusable.It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv. Use the --root-user-action option if you know what you are doing and want to suppress this warning.
(base) root@VM-24-95-ubuntu:/workspace# python -c "import tensorrt;print(tensorrt.__version__);assert tensorrt.Builder(tensorrt.Logger())"
8.5.3.1
[03/11/2025-02:03:52] [TRT] [W] CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#env-vars
Troubleshooting
1 nvidia驱动搞坏了
问题:
经过一系列的配置和安装,好像搞坏了哪里
原因:
- 可能是使用apt install, 或者bash NVIDIA-Linux-x86_64-XXX.XXX.XXX.run更新了驱动或者CUDA toolkit.然而这样更新驱动在Cloud Studio是不能成功更新的.
- 使用pip install应该不会把驱动环境搞坏
- 由于Cloud Studio的nvidia driver是以只读方式mount在容器空间中的,所以卸载掉用户安装的驱动即可恢复使用本来的驱动.(注意如果用户修改过$PATH或LD_LIBRARY_PATH环境变量,也需要恢复到原来的环境变量)
解决:apt remote *nvidia* -y
附件：
附件1:使用apt install更新驱动之后nvidia-smi错误
(base) root@VM-24-95-ubuntu:/workspace# apt install nvidia-driver-535 -y
WARNING: apt does not have a stable CLI interface. Use with caution in scripts.
Reading package lists...
Building dependency tree...
Reading state information...
The following additional packages will be installed:
libnvidia-cfg1-535 libnvidia-common-535 libnvidia-compute-535
libnvidia-decode-535 libnvidia-encode-535 libnvidia-extra-535
libnvidia-fbc1-535 libnvidia-gl-535 nvidia-compute-utils-535 nvidia-dkms-535
nvidia-kernel-common-535 nvidia-kernel-source-535 nvidia-prime
nvidia-settings nvidia-utils-535 xserver-xorg-video-nvidia-535
Recommended packages:
libnvidia-compute-535:i386 libnvidia-decode-535:i386
libnvidia-encode-535:i386 libnvidia-fbc1-535:i386 libnvidia-gl-535:i386
The following NEW packages will be installed:
libnvidia-cfg1-535 libnvidia-common-535 libnvidia-compute-535
libnvidia-decode-535 libnvidia-encode-535 libnvidia-extra-535
libnvidia-fbc1-535 libnvidia-gl-535 nvidia-compute-utils-535 nvidia-dkms-535
nvidia-driver-535 nvidia-kernel-common-535 nvidia-kernel-source-535
nvidia-prime nvidia-settings nvidia-utils-535 xserver-xorg-video-nvidia-535
0 upgraded, 17 newly installed, 0 to remove and 4 not upgraded.
Need to get 308 MB of archives.
After this operation, 801 MB of additional disk space will be used.
Get:1 http://mirrors.cloud.tencent.com/ubuntu focal-updates/main amd64 nvidia-prime all 0.8.16~0.20.04.2 [9960 B]
Get:2 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 libnvidia-cfg1-535 535.230.02-0ubuntu1 [98.9 kB]
Get:3 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 libnvidia-common-535 535.230.02-0ubuntu1 [14.9 kB]
Get:4 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 libnvidia-compute-535 535.230.02-0ubuntu1 [36.9 MB]
Get:5 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 libnvidia-decode-535 535.230.02-0ubuntu1 [1660 kB]
Get:6 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 libnvidia-encode-535 535.230.02-0ubuntu1 [90.0 kB]
Get:7 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 libnvidia-extra-535 535.230.02-0ubuntu1 [256 kB]
Get:8 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 libnvidia-fbc1-535 535.230.02-0ubuntu1 [51.3 kB]
Get:9 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 libnvidia-gl-535 535.230.02-0ubuntu1 [183 MB]
Get:10 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 nvidia-compute-utils-535 535.230.02-0ubuntu1 [285 kB]
Get:11 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 nvidia-kernel-source-535 535.230.02-0ubuntu1 [44.5 MB]
Get:12 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 nvidia-kernel-common-535 535.230.02-0ubuntu1 [38.4 MB]
Get:13 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 nvidia-dkms-535 535.230.02-0ubuntu1 [34.2 kB]
Get:14 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 nvidia-utils-535 535.230.02-0ubuntu1 [382 kB]
Get:15 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 xserver-xorg-video-nvidia-535 535.230.02-0ubuntu1 [1504 kB]
Get:16 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 nvidia-driver-535 535.230.02-0ubuntu1 [478 kB]
Get:17 https://developer.download.nvidia.cn/compute/cuda/repos/ubuntu2004/x86_64 nvidia-settings 570.124.06-0ubuntu1 [951 kB]
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
LANGUAGE = (unset),
LC_ALL = (unset),
LC_CTYPE = "C.UTF-8",
LANG = "en_US.UTF-8"
are supported and installed on your system.
perl: warning: Falling back to the standard locale ("C").
locale: Cannot set LC_MESSAGES to default locale: No such file or directory
locale: Cannot set LC_ALL to default locale: No such file or directory
debconf: delaying package configuration, since apt-utils is not installed
Fetched 308 MB in 27s (11.5 MB/s)
Selecting previously unselected package libnvidia-cfg1-535:amd64.
(Reading database ...
(Reading database ... 5%
(Reading database ... 10%
(Reading database ... 15%
(Reading database ... 20%
(Reading database ... 25%
(Reading database ... 30%
(Reading database ... 35%
(Reading database ... 40%
(Reading database ... 45%
(Reading database ... 50%
(Reading database ... 55%
(Reading database ... 60%
(Reading database ... 65%
(Reading database ... 70%
(Reading database ... 75%
(Reading database ... 80%
(Reading database ... 85%
(Reading database ... 90%
(Reading database ... 95%
(Reading database ... 100%
(Reading database ... 84774 files and directories currently installed.)
Preparing to unpack .../00-libnvidia-cfg1-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking libnvidia-cfg1-535:amd64 (535.230.02-0ubuntu1) ...
Selecting previously unselected package libnvidia-common-535.
Preparing to unpack .../01-libnvidia-common-535_535.230.02-0ubuntu1_all.deb ...
Unpacking libnvidia-common-535 (535.230.02-0ubuntu1) ...
Selecting previously unselected package libnvidia-compute-535:amd64.
Preparing to unpack .../02-libnvidia-compute-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking libnvidia-compute-535:amd64 (535.230.02-0ubuntu1) ...
Selecting previously unselected package libnvidia-decode-535:amd64.
Preparing to unpack .../03-libnvidia-decode-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking libnvidia-decode-535:amd64 (535.230.02-0ubuntu1) ...
Selecting previously unselected package libnvidia-encode-535:amd64.
Preparing to unpack .../04-libnvidia-encode-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking libnvidia-encode-535:amd64 (535.230.02-0ubuntu1) ...
Selecting previously unselected package libnvidia-extra-535:amd64.
Preparing to unpack .../05-libnvidia-extra-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking libnvidia-extra-535:amd64 (535.230.02-0ubuntu1) ...
Selecting previously unselected package libnvidia-fbc1-535:amd64.
Preparing to unpack .../06-libnvidia-fbc1-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking libnvidia-fbc1-535:amd64 (535.230.02-0ubuntu1) ...
Selecting previously unselected package libnvidia-gl-535:amd64.
Preparing to unpack .../07-libnvidia-gl-535_535.230.02-0ubuntu1_amd64.deb ...
dpkg-query: no packages found matching libnvidia-gl-450
Unpacking libnvidia-gl-535:amd64 (535.230.02-0ubuntu1) ...
Preparing to unpack .../08-nvidia-compute-utils-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking nvidia-compute-utils-535 (535.230.02-0ubuntu1) ...
dpkg: error processing archive /tmp/apt-dpkg-install-weWcQR/08-nvidia-compute-utils-535_535.230.02-0ubuntu1_amd64.deb (--unpack):
unable to make backup link of './usr/bin/nvidia-cuda-mps-control' before installing new version: Invalid cross-device link
dpkg-deb: error: paste subprocess was killed by signal (Broken pipe)
Selecting previously unselected package nvidia-kernel-source-535.
Preparing to unpack .../09-nvidia-kernel-source-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking nvidia-kernel-source-535 (535.230.02-0ubuntu1) ...
Selecting previously unselected package nvidia-kernel-common-535.
Preparing to unpack .../10-nvidia-kernel-common-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking nvidia-kernel-common-535 (535.230.02-0ubuntu1) ...
Selecting previously unselected package nvidia-dkms-535.
Preparing to unpack .../11-nvidia-dkms-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking nvidia-dkms-535 (535.230.02-0ubuntu1) ...
Preparing to unpack .../12-nvidia-utils-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking nvidia-utils-535 (535.230.02-0ubuntu1) ...
dpkg: error processing archive /tmp/apt-dpkg-install-weWcQR/12-nvidia-utils-535_535.230.02-0ubuntu1_amd64.deb (--unpack):
unable to make backup link of './usr/bin/nvidia-debugdump' before installing new version: Invalid cross-device link
dpkg-deb: error: paste subprocess was killed by signal (Broken pipe)
Selecting previously unselected package xserver-xorg-video-nvidia-535.
Preparing to unpack .../13-xserver-xorg-video-nvidia-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking xserver-xorg-video-nvidia-535 (535.230.02-0ubuntu1) ...
Selecting previously unselected package nvidia-driver-535.
Preparing to unpack .../14-nvidia-driver-535_535.230.02-0ubuntu1_amd64.deb ...
Unpacking nvidia-driver-535 (535.230.02-0ubuntu1) ...
Selecting previously unselected package nvidia-prime.
Preparing to unpack .../15-nvidia-prime_0.8.16~0.20.04.2_all.deb ...
Unpacking nvidia-prime (0.8.16~0.20.04.2) ...
Selecting previously unselected package nvidia-settings.
Preparing to unpack .../16-nvidia-settings_570.124.06-0ubuntu1_amd64.deb ...
Unpacking nvidia-settings (570.124.06-0ubuntu1) ...
Errors were encountered while processing:
/tmp/apt-dpkg-install-weWcQR/08-nvidia-compute-utils-535_535.230.02-0ubuntu1_amd64.deb
/tmp/apt-dpkg-install-weWcQR/12-nvidia-utils-535_535.230.02-0ubuntu1_amd64.deb
E: Sub-process /usr/bin/dpkg returned an error code (1)
(base) root@VM-24-95-ubuntu:/workspace# nvidia-smi
Failed to initialize NVML: Driver/library version mismatc10附件2:使用apt remove修复后nvidia-smi正常了
(base) root@VM-24-95-ubuntu:/workspace# apt remove *nvidia* -y
WARNING: apt does not have a stable CLI interface. Use with caution in scripts.
Reading package lists...
Building dependency tree...
Reading state information...
Package 'nvidia-304' is not installed, so not removed
这里省略了一些log
Package 'linux-objects-nvidia-535-server-5.15.0-1049-aws' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1049-azure' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1049-gcp' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1049-intel-iotg' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1049-oracle' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-105-generic' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-105-lowlatency' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1050-aws' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1050-azure' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1050-intel-iotg' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1050-oracle' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1051-aws' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1051-azure' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1051-gcp' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1051-oracle' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1052-aws' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1052-azure' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1052-gcp' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1052-intel-iotg' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1052-oracle' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1053-aws' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1053-azure' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1053-gcp' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1053-oracle' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1054-azure' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1055-aws' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1055-gcp' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1055-intel-iotg' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1055-oracle' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1056-azure' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1057-aws' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1057-azure' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1058-aws' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1058-azure' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1058-gcp' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1058-intel-iotg' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1058-oracle' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1059-gcp' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1059-intel-iotg' is not installed, so not removed
Package 'linux-objects-nvidia-535-server-5.15.0-1059-oracle' is not installed,
The following packages were automatically installed and are no longer required:
accountsservice acl apg apport apport-symptoms aptdaemon aptdaemon-data
aspell aspell-en avahi-daemon avahi-utils bind9-host bind9-libs bluez bolt
bubblewrap cheese-common colord colord-data cracklib-runtime crda
cups-pk-helper dconf-cli dctrl-tools desktop-file-utils dictionaries-common
dkms dns-root-data dnsmasq-base docbook-xml emacsen-common enchant-2
evolution-data-server evolution-data-server-common fprintd gdm3 geoclue-2.0
gettext-base gir1.2-accountsservice-1.0 gir1.2-atk-1.0 gir1.2-atspi-2.0
gir1.2-freedesktop gir1.2-gck-1 gir1.2-gcr-3 gir1.2-gdesktopenums-3.0
gir1.2-gdkpixbuf-2.0 gir1.2-gdm-1.0 gir1.2-geoclue-2.0
gir1.2-gnomebluetooth-1.0 gir1.2-gnomedesktop-3.0 gir1.2-graphene-1.0
gir1.2-gtk-3.0 gir1.2-gweather-3.0 gir1.2-ibus-1.0 gir1.2-json-1.0
gir1.2-mutter-6 gir1.2-nm-1.0 gir1.2-nma-1.0 gir1.2-notify-0.7
gir1.2-pango-1.0 gir1.2-polkit-1.0 gir1.2-rsvg-2.0 gir1.2-secret-1
gir1.2-soup-2.4 gir1.2-upowerglib-1.0 gir1.2-vte-2.91 gjs gkbd-capplet
gnome-control-center gnome-control-center-data gnome-control-center-faces
gnome-desktop3-data gnome-keyring gnome-keyring-pkcs11 gnome-menus
gnome-online-accounts gnome-session-bin gnome-session-common
gnome-settings-daemon gnome-settings-daemon-common gnome-shell
gnome-shell-common gnome-startup-applications gnome-user-docs groff-base
gstreamer1.0-clutter-3.0 gstreamer1.0-gl gstreamer1.0-plugins-base
gstreamer1.0-plugins-good gstreamer1.0-pulseaudio gstreamer1.0-x
hunspell-en-us ibus ibus-data ibus-gtk ibus-gtk3 iio-sensor-proxy im-config
ippusbxd iptables iw keyboard-configuration kmod language-selector-common
language-selector-gnome libaa1 libaccountsservice0 libappindicator3-1
libarchive13 libasound2-plugins libaspell15 libasyncns0 libavahi-core7
libavahi-glib1 libavc1394-0 libbluetooth3 libboost-thread1.71.0 libcaca0
libcamel-1.2-62 libcanberra-gtk3-0 libcanberra-gtk3-module libcanberra-pulse
libcdparanoia0 libcheese-gtk25 libcheese8 libclutter-1.0-0
libclutter-1.0-common libclutter-gst-3.0-0 libclutter-gtk-1.0-0
libcogl-common libcogl-pango20 libcogl-path20 libcogl20 libcolord-gtk1
libcolorhug2 libcrack2 libdaemon0 libdbusmenu-glib4 libdbusmenu-gtk3-4
libdrm-amdgpu1 libdrm-common libdrm-intel1 libdrm-nouveau2 libdrm-radeon1
libdrm2 libdv4 libebackend-1.2-10 libebook-1.2-20 libebook-contacts-1.2-3
libecal-2.0-1 libedata-book-1.2-26 libedata-cal-2.0-1 libedataserver-1.2-24
libedataserverui-1.2-2 libegl-mesa0 libegl1 libenchant-2-2 libevdev2
libexif12 libflac8 libfontenc1 libfprint-2-2 libgail-common libgail18
libgbm1 libgd3 libgdata-common libgdata22 libgdm1 libgee-0.8-2
libgeoclue-2-0 libgeocode-glib0 libgjs0g libgl1 libgl1-mesa-dri
libglapi-mesa libgles2 libglvnd0 libglx-mesa0 libglx0 libgnome-autoar-0-0
libgnome-bluetooth13 libgnome-desktop-3-19 libgnomekbd-common libgnomekbd8
libgoa-1.0-0b libgoa-1.0-common libgoa-backend-1.0-1 libgphoto2-6
libgphoto2-l10n libgphoto2-port12 libgraphene-1.0-0 libgsound0
libgssdp-1.2-0 libgstreamer-gl1.0-0 libgstreamer-plugins-base1.0-0
libgstreamer-plugins-good1.0-0 libgtk2.0-0 libgtk2.0-bin libgtk2.0-common
libgtop-2.0-11 libgtop2-common libgudev-1.0-0 libgupnp-1.2-0
libgupnp-av-1.0-2 libgupnp-dlna-2.0-3 libgusb2 libgweather-3-16
libgweather-common libharfbuzz-icu0 libhunspell-1.7-0 libhyphen0
libibus-1.0-5 libical3 libice6 libidn11 libiec61883-0 libieee1284-3
libimobiledevice6 libinput-bin libinput10 libip6tc2 libjack-jackd2-0
libjansson4 libjavascriptcoregtk-4.0-18 libldb2 libllvm12 libmaxminddb0
libmbim-glib4 libmbim-proxy libmediaart-2.0-0 libmm-glib0 libmnl0
libmozjs-68-0 libmp3lame0 libmpg123-0 libmtdev1 libmutter-6-0
libmysqlclient21 libndp0 libnetfilter-conntrack3 libnewt0.52 libnfnetlink0
libnftnl11 libnl-3-200 libnl-genl-3-200 libnl-route-3-200 libnm0 libnma0
libnotify4 libnspr4 libnss-mdns libnss3 libopengl0 libopus0 liborc-0.4-0
libpam-fprintd libpam-gnome-keyring libpangoxft-1.0-0 libpcap0.8 libpci3
libpciaccess0 libpcsclite1 libphonenumber7 libpipeline1 libplist3
libprotobuf17 libpulse-mainloop-glib0 libpulse0 libpulsedsp
libpwquality-common libpwquality1 libqmi-glib5 libqmi-proxy libraw1394-11
librygel-core-2.6-2 librygel-db-2.6-2 librygel-renderer-2.6-2
librygel-server-2.6-2 libsamplerate0 libsane libsane-common libsbc1
libsensors-config libsensors5 libshout3 libslang2 libsm6 libsmbclient
libsnapd-glib1 libsndfile1 libsnmp-base libsnmp35 libsodium23 libsoxr0
libspeex1 libspeexdsp1 libstartup-notification0 libtag1v5 libtag1v5-vanilla
libtalloc2 libteamdctl0 libtevent0 libtext-iconv-perl libtheora0 libtwolame0
libuchardet0 libudisks2-0 libunwind8 libupower-glib3 libusb-1.0-0
libusbmuxd6 libuv1 libv4l-0 libv4lconvert0 libvdpau1 libvisual-0.4-0
libvorbisenc2 libvpx6 libvte-2.91-0 libvte-2.91-common libvulkan1
libwacom-bin libwacom-common libwacom2 libwavpack1 libwayland-server0
libwbclient0 libwebkit2gtk-4.0-37 libwebpdemux2 libwebrtc-audio-processing1
libwhoopsie-preferences0 libwhoopsie0 libwoff1 libx11-xcb1 libxatracker2
libxaw7 libxcb-dri2-0 libxcb-dri3-0 libxcb-glx0 libxcb-icccm4 libxcb-image0
libxcb-keysyms1 libxcb-present0 libxcb-randr0 libxcb-render-util0
libxcb-res0 libxcb-shape0 libxcb-sync1 libxcb-util1 libxcb-xfixes0
libxcb-xkb1 libxcb-xv0 libxfont2 libxft2 libxkbcommon-x11-0 libxkbfile1
libxklavier16 libxmu6 libxnvctrl0 libxpm4 libxshmfence1 libxslt1.1 libxss1
libxt6 libxtables12 libxv1 libxvmc1 libxxf86vm1 libyelp0
linux-headers-5.4.0-208 linux-headers-5.4.0-208-generic
linux-headers-generic man-db mesa-vdpau-drivers mesa-vulkan-drivers
mobile-broadband-provider-info modemmanager mousetweaks mutter mutter-common
mysql-common network-manager network-manager-gnome network-manager-pptp
p11-kit p11-kit-modules pci.ids pkg-config ppp pptp-linux pulseaudio
pulseaudio-module-bluetooth pulseaudio-utils python3-apport
python3-aptdaemon python3-aptdaemon.gtk3widgets python3-blinker
python3-cairo python3-cffi-backend python3-cryptography python3-cups
python3-cupshelpers python3-defer python3-entrypoints python3-httplib2
python3-ibus-1.0 python3-jwt python3-keyring python3-launchpadlib
python3-lazr.restfulclient python3-lazr.uri python3-ldb
python3-macaroonbakery python3-nacl python3-oauthlib python3-problem-report
python3-protobuf python3-pymacaroons python3-rfc3339 python3-secretstorage
python3-simplejson python3-systemd python3-talloc python3-tz python3-wadllib
python3-xkit rtkit rygel samba-libs sane-utils screen-resolution-extra
session-migration sgml-base sgml-data sudo switcheroo-control
system-config-printer system-config-printer-common
system-config-printer-udev ubuntu-docs ubuntu-session ubuntu-wallpapers
ubuntu-wallpapers-focal udev update-inetd upower usb-modeswitch
usb-modeswitch-data usb.ids usbmuxd vdpau-driver-all wamerican
whoopsie-preferences wireless-regdb wpasupplicant x11-xkb-utils
x11-xserver-utils xdg-dbus-proxy xfonts-base xfonts-encodings xfonts-utils
xml-core xserver-common xserver-xephyr xserver-xorg xserver-xorg-core
xserver-xorg-input-all xserver-xorg-input-libinput xserver-xorg-input-wacom
xserver-xorg-legacy xserver-xorg-video-all xserver-xorg-video-amdgpu
xserver-xorg-video-ati xserver-xorg-video-fbdev xserver-xorg-video-intel
xserver-xorg-video-nouveau xserver-xorg-video-qxl xserver-xorg-video-radeon
xserver-xorg-video-vesa xserver-xorg-video-vmware xwayland
yaru-theme-gnome-shell yelp yelp-xsl zenity zenity-common
Use 'apt autoremove' to remove them.
The following packages will be REMOVED:
libnvidia-cfg1-535 libnvidia-common-535 libnvidia-compute-535
libnvidia-decode-535 libnvidia-encode-535 libnvidia-extra-535
libnvidia-fbc1-535 libnvidia-gl-535 nvidia-dkms-535 nvidia-driver-535
nvidia-kernel-common-535 nvidia-kernel-source-535 nvidia-prime
nvidia-settings xserver-xorg-video-nvidia-535
perl: warning: Setting locale failed.
perl: warning: Please check that your locale settings:
LANGUAGE = (unset),
LC_ALL = (unset),
LC_CTYPE = "C.UTF-8",
LANG = "en_US.UTF-8"
are supported and installed on your system.
perl: warning: Falling back to the standard locale ("C").
locale: Cannot set LC_MESSAGES to default locale: No such file or directory
locale: Cannot set LC_ALL to default locale: No such file or directory
0 upgraded, 0 newly installed, 15 to remove and 4 not upgraded.
15 not fully installed or removed.
After this operation, 798 MB disk space will be freed.
(Reading database ...
(Reading database ... 5%
(Reading database ... 10%
(Reading database ... 15%
(Reading database ... 20%
(Reading database ... 25%
(Reading database ... 30%
(Reading database ... 35%
(Reading database ... 40%
(Reading database ... 45%
(Reading database ... 50%
(Reading database ... 55%
(Reading database ... 60%
(Reading database ... 65%
(Reading database ... 70%
(Reading database ... 75%
(Reading database ... 80%
(Reading database ... 85%
(Reading database ... 90%
(Reading database ... 95%
(Reading database ... 100%
(Reading database ... 85474 files and directories currently installed.)
Removing nvidia-driver-535 (535.230.02-0ubuntu1) ...
Removing xserver-xorg-video-nvidia-535 (535.230.02-0ubuntu1) ...
Removing libnvidia-cfg1-535:amd64 (535.230.02-0ubuntu1) ...
Removing libnvidia-gl-535:amd64 (535.230.02-0ubuntu1) ...
Removing libnvidia-common-535 (535.230.02-0ubuntu1) ...
Removing libnvidia-encode-535:amd64 (535.230.02-0ubuntu1) ...
Removing libnvidia-decode-535:amd64 (535.230.02-0ubuntu1) ...
Removing libnvidia-compute-535:amd64 (535.230.02-0ubuntu1) ...
Removing libnvidia-extra-535:amd64 (535.230.02-0ubuntu1) ...
Removing libnvidia-fbc1-535:amd64 (535.230.02-0ubuntu1) ...
Removing nvidia-dkms-535 (535.230.02-0ubuntu1) ...
Removing nvidia-kernel-common-535 (535.230.02-0ubuntu1) ...
Removing nvidia-kernel-source-535 (535.230.02-0ubuntu1) ...
Removing nvidia-prime (0.8.16~0.20.04.2) ...
Removing nvidia-settings (570.124.06-0ubuntu1) ...
Processing triggers for mime-support (3.64ubuntu1) ...
Processing triggers for gnome-menus (3.36.0-1ubuntu1) ...
Processing triggers for libc-bin (2.31-0ubuntu9.17) ...
Processing triggers for man-db (2.9.1-1) ...
Processing triggers for dbus (1.12.16-2ubuntu2.3) ...
Processing triggers for desktop-file-utils (0.24-1ubuntu3) ...
(base) root@VM-24-95-ubuntu:/workspace# nvidia-smi
Fri Mar 14 03:24:46 2025
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.105.17 Driver Version: 525.105.17 CUDA Version: 12.0 |
|-------------------------------+----------------------+----------------------+
| 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 T4 On | 00000000:00:09.0 Off | 0 |
| N/A 31C P8 9W / 70W | 2MiB / 15360MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+本文参与 腾讯云自媒体同步曝光计划，分享自微信公众号。
原始发表：2025-03-14，如有侵权请联系 cloudcommunity@tencent.com 删除
本文分享自 腾讯云CloudStudio 微信公众号，前往查看
如有侵权，请联系 cloudcommunity@tencent.com 删除。
本文参与 腾讯云自媒体同步曝光计划 ，欢迎热爱写作的你一起参与！
评论
登录后参与评论
推荐阅读
腾讯云开发者
