解决win10下CuDNN报错：Loaded runtime CuDNN library: 7.5.0 but source was compiled with:

解决win10下CuDNN报错：Loaded runtime CuDNN library: 7.5.0 but source was compiled with:

背景描述

CUDNN的正确调用，关系着系统能否正常运行。
本人在Eclipse上有个工程项目，一直运转良好，今天突发情况，开始报错。“Loaded runtime CuDNN library: 7.5.0 but source was compiled with: 7.6.0”

网上关于这个问题的帖子有几个，但是都没能解决我的问题。后自行研究，并将其解决。为了帮助更多人避过此坑，故有此文

关联问题

如果你第一次面对该问题，可以先参考如下帖子。如果能解决最好，如果解决不了，可以继续看下面的内容

https://blog.csdn.net/caicaiyunyun/article/details/105334923
https://blog.csdn.net/weixin_48624971/article/details/107652341
https://stackoverflow.com/questions/61039316/loaded-runtime-cudnn-library-7-1-2-but-source-was-compiled-with-7-6-0-ubuntu

新的解决方案：思路分析

1、确实下载了正确版本（7.6以上）的CUDNN
2、确实覆盖了正确的位置C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0
3、确实更新了环境变量4条（4件套）
4、确实被报错了：“Loaded runtime CuDNN library: 7.5.0”，后边的信息与TF或是其他组件版本有关（可能不同）

新的解决方案：依赖级别查找

提问：Anaconda中的其他开发包中，是否也包括了Cudnn文件呢？
经过研究发现，却有开发包中包含了Cudnn文件，以本项目为例，MXNET、PyTorch中都有cudnn64_7.dll文件。（接近400M的东西多次重复出现，真是令人始料未及）
而这些文件的调用顺序，有可能优先于环境变量中指示的文件。

新的解决方案：解决思路

发现了如上现象，不可不作为。
MXNET与Pytorch中，都还有该文件（重要者？悲哀者？）
其属性如下（主要关注产品名称的版本号）

而我们从NV下载的新鲜文件，属性是这样的

新的解决方案：找到头绪，胆大心细，开干

将属性不同的入侵者全都重命名、备份，然后替换。
重新跑起咱们的项目，问题解决！

成功的一瞬间：开心

pciBusID: 0000:01:00.0 name: GeForce RTX 2080 Ti computeCapability: 7.5
coreClock: 1.635GHz coreCount: 68 deviceMemorySize: 11.00GiB deviceMemoryBandwidth: 573.69GiB/s
2020-08-27 15:33:18.479565: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudart64_101.dll
2020-08-27 15:33:18.479717: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cublas64_10.dll
2020-08-27 15:33:18.479869: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cufft64_10.dll
2020-08-27 15:33:18.480014: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library curand64_10.dll
2020-08-27 15:33:18.480194: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cusolver64_10.dll
2020-08-27 15:33:18.480341: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cusparse64_10.dll
2020-08-27 15:33:18.480488: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudnn64_7.dll
2020-08-27 15:33:18.480849: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1703] Adding visible gpu devices: 0
2020-08-27 15:33:18.974363: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1102] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-08-27 15:33:18.974544: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1108]      0 
2020-08-27 15:33:18.974644: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1121] 0:   N 
2020-08-27 15:33:18.975212: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1247] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 8590 MB memory) -> physical GPU (device: 0, name: GeForce RTX 2080 Ti, pci bus id: 0000:01:00.0, compute capability: 7.5)
2020-08-27 15:33:18.976801: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x154c5192e60 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:
2020-08-27 15:33:18.976994: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): GeForce RTX 2080 Ti, Compute Capability 7.5
processing: TempJPG.jpg
2020-08-27 15:33:20.176386: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudnn64_7.dll
2020-08-27 15:33:21.291192: W tensorflow/stream_executor/gpu/redzone_allocator.cc:314] Internal: Invoking GPU asm compilation is supported on Cuda non-Windows platforms only
Relying on driver to perform ptx compilation. 
Modify $PATH to customize ptxas location.
This message will be only logged once.
2020-08-27 15:33:21.383042: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cublas64_10.dll
All file processed
totally processed:1files
totally passed:0 files