3. cuDNN 8.0.4 설치합니다.

아래 파일 다운로두 후 설치 :

https://developer.nvidia.com/compute/machine-learning/cudnn/secure/8.0.4/11.0_20200923/cudnn-11.0-windows-x64-v8.0.4.30.zip

4. VSCode 에서 GPU 버젼이 제대로 설치 되었는지 확인합니다.

로그는 다음과 비슷하게 나오면 설치 완료 된것입니다.

참고로 저는 Nvidia GTX1080을 사용했습니다.

gputest.py

import tensorflow as tf

tf.debugging.set_log_device_placement(True)

​

# Create some tensors

a = tf.constant([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]])

b = tf.constant([[1.0, 2.0], [3.0, 4.0], [5.0, 6.0]])

c = tf.matmul(a, b)

print(c)

​

​

출력 로그:

2021-01-06 21:22:59.401338: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudart64_110.dll

2021-01-06 21:23:02.153274: I tensorflow/compiler/jit/xla_cpu_device.cc:41] Not creating XLA devices, tf_xla_enable_xla_devices not set

2021-01-06 21:23:02.154374: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library nvcuda.dll

2021-01-06 21:23:02.173672: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1720] Found device 0 with properties:

pciBusID: 0000:01:00.0 name: GeForce GTX 1080 computeCapability: 6.1

coreClock: 1.8475GHz coreCount: 20 deviceMemorySize: 8.00GiB deviceMemoryBandwidth: 298.32GiB/s

2021-01-06 21:23:02.173788: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudart64_110.dll

2021-01-06 21:23:02.180652: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublas64_11.dll

2021-01-06 21:23:02.180747: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublasLt64_11.dll

2021-01-06 21:23:02.184686: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cufft64_10.dll

2021-01-06 21:23:02.186262: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library curand64_10.dll

2021-01-06 21:23:02.194478: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusolver64_10.dll

2021-01-06 21:23:02.197588: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusparse64_11.dll

2021-01-06 21:23:02.198484: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudnn64_8.dll

2021-01-06 21:23:02.198858: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1862] Adding visible gpu devices: 0

2021-01-06 21:23:02.199563: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: AVX2

To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.

2021-01-06 21:23:02.201289: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1720] Found device 0 with properties:

pciBusID: 0000:01:00.0 name: GeForce GTX 1080 computeCapability: 6.1

coreClock: 1.8475GHz coreCount: 20 deviceMemorySize: 8.00GiB deviceMemoryBandwidth: 298.32GiB/s

2021-01-06 21:23:02.202097: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudart64_110.dll

2021-01-06 21:23:02.202355: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublas64_11.dll

2021-01-06 21:23:02.202669: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublasLt64_11.dll

2021-01-06 21:23:02.203360: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cufft64_10.dll

2021-01-06 21:23:02.203771: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library curand64_10.dll

2021-01-06 21:23:02.204080: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusolver64_10.dll

2021-01-06 21:23:02.204429: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cusparse64_11.dll

2021-01-06 21:23:02.204839: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cudnn64_8.dll

2021-01-06 21:23:02.205146: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1862] Adding visible gpu devices: 0

2021-01-06 21:23:02.729407: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1261] Device interconnect StreamExecutor with strength 1 edge matrix:

2021-01-06 21:23:02.729750: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1267] 0

2021-01-06 21:23:02.730221: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1280] 0: N

2021-01-06 21:23:02.730619: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1406] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 6692 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1080, pci bus id: 0000:01:00.0, compute capability: 6.1)

2021-01-06 21:23:02.745170: I tensorflow/compiler/jit/xla_gpu_device.cc:99] Not creating XLA devices, tf_xla_enable_xla_devices not set

2021-01-06 21:23:02.747587: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublas64_11.dll

2021-01-06 21:23:03.000677: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library cublasLt64_11.dll

2021-01-06 21:23:03.256567: I tensorflow/core/common_runtime/eager/execute.cc:760] Executing op MatMul in device /job:localhost/replica:0/task:0/device:GPU:0

tf.Tensor(

[[22. 28.]

[49. 64.]], shape=(2, 2), dtype=float32)

​

​

마지막 로그에서

2021-01-06 21:23:03.256567: I tensorflow/core/common_runtime/eager/execute.cc:760] Executing op MatMul in device /job:localhost/replica:0/task:0/device:GPU:0

--> GPU:0 처럼 설정되어 있다면 GPU로 계산이 제대로 된 상태입니다.

​

한가지 더 GPU 사용량을 보시면 더 확실히 확인 할 수 있습니다.