cuDNN下載安裝教學

- 10月 31, 2026

https://brian-liu.medium.com/build-enviroment-cuda-cudnn-on-win11-6ac5e51268f3

接續前一篇

關於GPU顯卡TFLOPS指標介紹和CUDA Toolkit下載安裝教學

裝完CUDA Toolkit之外還需要下載安裝cuDNN

https://developer.nvidia.com/rdp/cudnn-archive

由於我這邊裝的版本是CUDA Toolkit 11.2.2

因此從最新版本cuDNN v8.9.7、cuDNN v8.9.5、.....只要有匹配的CUDA 11.x都可以下載安裝

不然預設你的本地電腦Tensorflow程式，是無法啟用到GPU資源的。

Python測試腳本

import tensorflow as tf

print("TensorFlow:", tf.__version__)

gpus = tf.config.list_physical_devices("GPU")
print("Detected GPU:", gpus)
print('GPU->',tf.test.is_gpu_available())
if gpus:
    try:
        for gpu in gpus:
            tf.config.experimental.set_memory_growth(gpu, True)

        with tf.device("/GPU:0"):
            a = tf.random.normal((512, 512))
            b = tf.matmul(a, a)

        print("GPU calculation:", b.device)
        print("GPU test: PASS")

    except Exception as e:
        print("GPU test: FAIL")
        print(e)
else:
    print("GPU test: No GPU detected")
#如果成功，你的輸出應類似：
# TensorFlow: 2.10.0
# Detected GPU: [PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]
# GPU calculation: /job:localhost/replica:0/task:0/device:GPU:0
# GPU test: PASS

完整錯誤訊息就是沒有識別到相關dll

2026-11-01 11:37:23.035351: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'cudart64_110.dll'; dlerror: cudart64_110.dll not found

2026-11-01 11:37:23.035895: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.

TensorFlow: 2.10.0

2026-11-01 11:37:29.470594: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'cudart64_110.dll'; dlerror: cudart64_110.dll not found

2026-11-01 11:37:29.475030: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'cublas64_11.dll'; dlerror: cublas64_11.dll not found

2026-11-01 11:37:29.479769: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'cublasLt64_11.dll'; dlerror: cublasLt64_11.dll not found

2026-11-01 11:37:29.484110: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'cufft64_10.dll'; dlerror: cufft64_10.dll not found

2026-11-01 11:37:29.488224: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'curand64_10.dll'; dlerror: curand64_10.dll not found

2026-11-01 11:37:29.492378: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'cusolver64_11.dll'; dlerror: cusolver64_11.dll not found

2026-11-01 11:37:29.497203: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'cusparse64_11.dll'; dlerror: cusparse64_11.dll not found

2026-11-01 11:37:29.501809: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'cudnn64_8.dll'; dlerror: cudnn64_8.dll not found

2026-11-01 11:37:29.501929: W tensorflow/core/common_runtime/gpu/gpu_device.cc:1934] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at https://www.tensorflow.org/install/gpu for how to download and setup the required libraries for your platform.

Skipping registering GPU devices...

這邊我選這版本