Using gpu vs cpu in tensorflow deep mnist example

Question

The program i am using i copy-pasted from here with a few changes. This is my code with an attempt to boost the speed of training:

from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets('MNIST_data', one_hot=True)

import tensorflow as tf

x = tf.placeholder(tf.float32, shape=[None, 784])
y_ = tf.placeholder(tf.float32, shape=[None, 10])
W = tf.Variable(tf.zeros([784,10]))
b = tf.Variable(tf.zeros([10]))

def weight_variable(shape):
  initial = tf.truncated_normal(shape, stddev=0.1)
  return tf.Variable(initial)

def bias_variable(shape):
  initial = tf.constant(0.1, shape=shape)
  return tf.Variable(initial)

def conv2d(x, W):
  return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')

def max_pool_2x2(x):
  return tf.nn.max_pool(x, ksize=[1, 2, 2, 1],
                        strides=[1, 2, 2, 1], padding='SAME')

with tf.device('/gpu:0'):
  W_conv1 = weight_variable([5, 5, 1, 32])
  b_conv1 = bias_variable([32])
  x_image = tf.reshape(x, [-1, 28, 28, 1])
  h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)
  h_pool1 = max_pool_2x2(h_conv1)

  W_conv2 = weight_variable([5, 5, 32, 64])
  b_conv2 = bias_variable([64])

  h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)
  h_pool2 = max_pool_2x2(h_conv2)

  W_fc1 = weight_variable([7 * 7 * 64, 1024])
  b_fc1 = bias_variable([1024])

  h_pool2_flat = tf.reshape(h_pool2, [-1, 7*7*64])
  h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)

  keep_prob = tf.placeholder(tf.float32)
  h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)

  W_fc2 = weight_variable([1024, 10])
  b_fc2 = bias_variable([10])

  y_conv = tf.matmul(h_fc1_drop, W_fc2) + b_fc2

  cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=y_, logits=y_conv))
  train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)
  correct_prediction = tf.equal(tf.argmax(y_conv, 1), tf.argmax(y_, 1))
  accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))

  with tf.Session(config=tf.ConfigProto(allow_soft_placement=True, log_device_placement=True)) as sess:
    sess.run(tf.global_variables_initializer())
    for i in range(20000):
      batch = mnist.train.next_batch(50)
      if i % 100 == 0:
        train_accuracy = accuracy.eval(feed_dict={
          x: batch[0], y_: batch[1], keep_prob: 1.0})
        print('step %d, training accuracy %g' % (i, train_accuracy))
      train_step.run(feed_dict={x: batch[0], y_: batch[1], keep_prob: 0.5})

  print('test accuracy %g' % accuracy.eval(feed_dict={
      x: mnist.test.images, y_: mnist.test.labels, keep_prob: 1.0}))

Which produces the following output:

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
step 0, training accuracy 0.22
step 100, training accuracy 0.76
step 200, training accuracy 0.88
...

The problem is that the time taken by the original code on the tutorial (ie without the with tf.device('/gpu:0'): on line 26) and this code have no measurable difference (about 10 seconds for each step). I have installed cuda-8.0 and cuDNN successfully (after many hours of failed attempts). "$ nvidia-smi" returns the following output:

Sun Jul  2 13:57:10 2017       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 375.26                 Driver Version: 375.26                    |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GT 710      Off  | 0000:01:00.0     N/A |                  N/A |
| N/A   49C    P0    N/A /  N/A |    406MiB /  2000MiB |     N/A      Default |
+-------------------------------+----------------------+----------------------+


+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID  Type  Process name                               Usage      |
|=============================================================================|
|    0                  Not Supported                                         |
+-----------------------------------------------------------------------------+

So the question is:

1) Is the work too small to produce no difference in choosing a cpu or a gpu? 2) Or there some stupid mistake in my implementation?

Thanks for reading the whole question.

Answer 1

The fact that you can run this code with no errors suggest that TensorFlow can definitely run with a GPU. The problem here is that when you run TensorFlow as is, by default, it tries to run on the GPU. There are a few ways you can force it to run on the CPU.

1. Run it this way: CUDA_VISIBLE_DEVICES= python code.py. Note that when you do this and still have with tf.device('/gpu:0'), it will break, so remove it.
2. Change with tf.device('/gpu:0') to with tf.device('/cpu:0')

EDIT from question in comments

See here for more information on what allow_soft_placement and log_device_placement mean in ConfigProto.