本博文是对前面两篇tensorflow的博文的一个继续，对环境的更新。
基于tensorflow的MNIST手写识别
安装tensorflow，那叫一个坑啊
主要出发点：
上述两篇博文的程序运行的环境，其实是没有用到GPU的。本篇博文，介绍如何利用GPU。
首先通过pip重新安装一个支持gpu的tensorflow，采用upgrade的方式进行。
[root@bogon tensorflow]# pip install --upgrade tensorflow-gpu
Collecting tensorflow-gpu
Downloading tensorflow_gpu-1.0.-cp27-cp27mu-manylinux1_x86_64.whl (.8MB)
% |████████████████████████████████| .8MB .6kB/s
Requirement already up-to-date: protobuf>=3.1. in /usr/lib64/python2./site-packages (from tensorflow-gpu)
Requirement already up-to-date: six>=1.10. in /usr/lib/python2./site-packages (from tensorflow-gpu)
Requirement already up-to-date: wheel in /usr/lib/python2./site-packages (from tensorflow-gpu)
Requirement already up-to-date: mock>=2.0. in /usr/lib/python2./site-packages (from tensorflow-gpu)
Requirement already up-to-date: numpy>=1.11. in /usr/lib64/python2./site-packages (from tensorflow-gpu)
Requirement already up-to-date: setuptools in /usr/lib/python2./site-packages (from protobuf>=3.1.->tensorflow-gpu)
Requirement already up-to-date: funcsigs>=; python_version < "3.3" in /usr/lib/python2./site-packages (from mock>=2.0.->tensorflow-gpu)
Requirement already up-to-date: pbr>=0.11 in /usr/lib/python2./site-packages (from mock>=2.0.->tensorflow-gpu)
Requirement already up-to-date: appdirs>=1.4. in /usr/lib/python2./site-packages (from setuptools->protobuf>=3.1.->tensorflow-gpu)
Requirement already up-to-date: packaging>=16.8 in /usr/lib/python2./site-packages (from setuptools->protobuf>=3.1.->tensorflow-gpu)
Requirement already up-to-date: pyparsing in /usr/lib/python2./site-packages (from packaging>=16.8->setuptools->protobuf>=3.1.->tensorflow-gpu)
Installing collected packages: tensorflow-gpu
Successfully installed tensorflow-gpu-1.0.
这个过程顺利完成。
然后，将MNIST的手写识别程序，在运行一下，验证一下，是否启用GPU。
[root@bogon tensorflow]# python mnist_demo1.py
I tensorflow/stream_executor/dso_loader.cc:] successfully opened CUDA library libcublas.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:126] Couldn't open CUDA library libcudnn.so.5. LD_LIBRARY_PATH: /usr/local/cuda-8.0/lib64:
I tensorflow/stream_executor/cuda/cuda_dnn.cc:3517] Unable to load cuDNN DSO
I tensorflow/stream_executor/dso_loader.cc:] successfully opened CUDA library libcufft.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:] successfully opened CUDA library libcuda.so. locally
I tensorflow/stream_executor/dso_loader.cc:] successfully opened CUDA library libcurand.so.8.0 locally
Extracting MNIST_data/train-images-idx3-ubyte.gz
Extracting MNIST_data/train-labels-idx1-ubyte.gz
Extracting MNIST_data/t10k-images-idx3-ubyte.gz
Extracting MNIST_data/t10k-labels-idx1-ubyte.gz
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use SSE3 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
I tensorflow/core/common_runtime/gpu/gpu_device.cc:] Found device with properties:
name: GeForce GTX
major: minor: memoryClockRate (GHz) 1.7335
pciBusID ::00.0
Total memory: .92GiB
Free memory: .81GiB
I tensorflow/core/common_runtime/gpu/gpu_device.cc:] DMA:
I tensorflow/core/common_runtime/gpu/gpu_device.cc:] : Y
I tensorflow/core/common_runtime/gpu/gpu_device.cc:] Creating TensorFlow device (/gpu:) -> (device: , name: GeForce GTX , pci bus id: ::00.0)
F tensorflow/stream_executor/cuda/cuda_dnn.cc:] Check failed: s.ok() could not find cudnnCreate in cudnn DSO; dlerror: /usr/lib/python2./site-packages/tensorflow/python/_pywrap_tensorflow.so: undefined symbol: cudnnCreate
Aborted (core dumped)
上面红色部分报错了，找不到cudnn的so文件，进入到cuda的安装路径，查看是否有这个so。
[root@bogon lib64]# ll libcudnn
libcudnn.so.5.1 libcudnn.so.5.1. libcudnn_static.a
的确没有libcudnn.so.5的文件。
下面，建立一个软连接，将libcudnn.so.5指向libcudnn.so.5.1。
[root@bogon lib64]# ln -s libcudnn.so.5.1 libcudnn.so.5
[root@bogon lib64]# ll libcudnn*
lrwxrwxrwx. root root Mar : libcudnn.so. -> libcudnn.so.5.1
lrwxrwxrwx. root root Mar : libcudnn.so.5.1 -> libcudnn.so.5.1.
-rwxr-xr-x. root root Mar : libcudnn.so.5.1.
-rw-r--r--. root root Mar : libcudnn_static.a
现在，有了这个libcudnn.so.5的文件了。
再次验证mnist的手写识别程序。
[root@bogon tensorflow]# python mnist_demo1.py
I tensorflow/stream_executor/dso_loader.cc:] successfully opened CUDA library libcublas.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:] successfully opened CUDA library libcudnn.so. locally
I tensorflow/stream_executor/dso_loader.cc:] successfully opened CUDA library libcufft.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:] successfully opened CUDA library libcuda.so. locally
I tensorflow/stream_executor/dso_loader.cc:] successfully opened CUDA library libcurand.so.8.0 locally
Extracting MNIST_data/train-images-idx3-ubyte.gz
Extracting MNIST_data/train-labels-idx1-ubyte.gz
Extracting MNIST_data/t10k-images-idx3-ubyte.gz
Extracting MNIST_data/t10k-labels-idx1-ubyte.gz
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use SSE3 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
I tensorflow/core/common_runtime/gpu/gpu_device.cc:] Found device with properties:
name: GeForce GTX
major: minor: memoryClockRate (GHz) 1.7335
pciBusID ::00.0
Total memory: .92GiB
Free memory: .81GiB
I tensorflow/core/common_runtime/gpu/gpu_device.cc:] DMA:
I tensorflow/core/common_runtime/gpu/gpu_device.cc:] : Y
I tensorflow/core/common_runtime/gpu/gpu_device.cc:] Creating TensorFlow device (/gpu:) -> (device: , name: GeForce GTX , pci bus id: ::00.0)
0.9092
到现在为止，我的tensorflow的运行环境，已经是基于GPU的了。
下面附上测试中的mnist_demo1.py的内容：
#!/usr/bin/env python
# -*- coding: utf- -*- import tensorflow as tf import tensorflow.examples.tutorials.mnist.input_data as input_data
mnist = input_data.read_data_sets("MNIST_data", one_hot=True) sess = tf.InteractiveSession() x = tf.placeholder("float", shape=[None, ])
y_ = tf.placeholder("float", shape=[None, ]) w = tf.Variable(tf.zeros([,]))
b = tf.Variable(tf.zeros([])) init = tf.global_variables_initializer()
sess.run(init) y = tf.nn.softmax(tf.matmul(x, w) + b) cross_entropy = -tf.reduce_sum(y_*tf.log(y)) train_step = tf.train.GradientDescentOptimizer(0.01).minimize(cross_entropy) for i in range():
batch = mnist.train.next_batch()
train_step.run(feed_dict={x: batch[], y_: batch[]}) correct_prediction = tf.equal(tf.argmax(y,), tf.argmax(y_,))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float")) print accuracy.eval(feed_dict={x: mnist.test.images, y_: mnist.test.labels})
最后说明下，上述WARNING部分：
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use SSE3 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
暂时没有关注，所知道的处理办法，就是用bazel进行源码安装tensorflow可以解决这个问题。由于不是太影响实验，暂且不关注。