前往小程序，Get更优阅读体验！
立即前往
文档建议反馈控制台
首页
学习
活动
专区
圈层
工具
MCP广场
文章/答案/技术大牛
发布
首页
学习
活动
专区
圈层
工具
MCP广场
返回腾讯云官网
社区首页 >专栏 >什么是cuDNN?如何安装CUDA和cuDNN

什么是cuDNN?如何安装CUDA和cuDNN

作者头像
CloudStudio
发布于 2025-03-13 19:34:01
发布于 2025-03-13 19:34:01
1K00
代码可运行
举报
文章被收录于专栏：《Cloud Studio》《Cloud Studio》
运行总次数：0
代码可运行

文章原文指路：https://juejin.cn/post/7479993915041660968

正文如下：

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

截屏2025-03-10 20.01.11.png
截屏2025-03-10 20.01.11.png
  • 选择免费基础版 -> 确认
截屏2025-03-10 20.02.48.png
截屏2025-03-10 20.02.48.png
  • 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.
截屏2025-03-10 20.03.57.png
截屏2025-03-10 20.03.57.png
  • 点击Pytorch2.0.0 gssrak进入空间,等待不到一分钟则会加载完成
截屏2025-03-10 20.06.39.png
截屏2025-03-10 20.06.39.png

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:
代码语言：txt
复制
(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
代码语言：txt
复制
(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_0
  • echo $PATH,检查是否包含过了路径/usr/local/cuda/bin
代码语言：txt
复制
(base) root@VM-24-95-ubuntu:/workspace# echo $PATH
/etc/.hai/cloud_studio/vendor/modules/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:/bin

  • echo $LD_LIBRARY_PATH,检查是否包含过了路径/usr/local/cuda/lib64
代码语言：txt
复制
 (base) root@VM-24-95-ubuntu:/workspace# echo $LD_LIBRARY_PATH
/usr/local/nvidia/lib:/usr/local/nvidia/lib64:/usr/local/cuda/lib64

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.

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cudapython -c "import torch;print(torch.cuda.is_available())"
  • 查看cuDNN是否启用python -c "import torch;print(torch.backends.cudnn.enabled)"
  • 查看cuDNN版本python -c "import torch;print(torch.backends.cudnn.version())"
代码语言：txt
复制
(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*"
代码语言：txt
复制
(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.8

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`

代码语言：javascript
代码运行次数：0
运行
复制
bash 代码解读复制代码(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
    • 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):conda install cudnn cuda-version=<cuda-major-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)

tensorRT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`

代码语言：javascript
代码运行次数：0
运行
复制
bash 代码解读复制代码(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-loading

logs of `pip install tensorrt==8.5.3.1`

代码语言：txt
复制
(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

本文系转载，前往查看

如有侵权，请联系 cloudcommunity@tencent.com 删除。

Cloud Studio（云端 IDE）

本文系转载，前往查看

如有侵权，请联系 cloudcommunity@tencent.com 删除。

Cloud Studio（云端 IDE）
评论
登录后参与评论
0 条评论
热度
最新
登录 后参与评论
推荐阅读
目录
  • cloud studio介绍
    • 开启Cloud Studio GPU空间
  • Nvidia driver
  • CUDA toolkit
  • cuDNN
    • cuDNN介绍
    • 查看cuDNN版本
      • 验证cuDNN安装
    • 手动安装/升级cuDNN(可选)
  • tensorRT(可选)
领券

腾讯云开发者

扫码关注腾讯云开发者

扫码关注腾讯云开发者

领取腾讯云代金券

热门产品

热门推荐

更多推荐

Copyright © 2013 - 2025 Tencent Cloud. All Rights Reserved. 腾讯云 版权所有 

深圳市腾讯计算机系统有限公司 ICP备案/许可证号：粤B2-20090059 深公网安备号 44030502008569

腾讯云计算（北京）有限责任公司 京ICP证150476号 |  京ICP备11018762号 | 京公网安备号11010802020287

问题归档专栏文章快讯文章归档关键词归档开发者手册归档开发者手册 Section 归档

Copyright © 2013 - 2025 Tencent Cloud.

All Rights Reserved. 腾讯云 版权所有

登录 后参与评论