想要测试Tensorflow-gpu装好了吗，run它，run出来那就可以，测试你的tensorflow是否和cuda和cudnn匹配，如果运行成功没有报错且你可以看到你的显卡名称。

import tensorflow as tf
import datetime

# running
# Creates a graph.(cpu version)
print('cpu version')
starttime1 = datetime.datetime.now()
with tf.device('/cpu:0'):
    a = tf.constant(
        [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0,
         5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0,
         3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[6, 9], name='a')
    b = tf.constant(
        [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0,
         5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0,
         3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[9, 6], name='b')
    c = tf.matmul(a, b)
    c = tf.matmul(c, a)
    c = tf.matmul(c, b)
# Creates a session with log_device_placement set to True.
sess1 = tf.Session(config=tf.ConfigProto(log_device_placement=True))
# Runs the op.
for i in range(10):
    sess1.run(c)
print(sess1.run(c))
sess1.close()
endtime1 = datetime.datetime.now()
time1 = (endtime1 - starttime1).microseconds
# print('time1:',time1)
#############################################
print('gpuversion')
# Creates a graph.(gpu version)
starttime2 = datetime.datetime.now()
# running
with tf.device('/gpu:0'):
    a = tf.constant(
        [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0,
         5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0,
         3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[6, 9], name='a')
    b = tf.constant(
        [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0,
         5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 1.0, 2.0,
         3.0, 4.0, 5.0, 6.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[9, 6], name='b')
    c = tf.matmul(a, b)
    c = tf.matmul(c, a)
    c = tf.matmul(c, b)
# Creates a session with log_device_placement set to True.
sess2 = tf.Session(config=tf.ConfigProto(log_device_placement=True))
# Runs the op.
for i in range(10):
    sess2.run(c)
print(sess2.run(c))
sess2.close()
endtime2 = datetime.datetime.now()
time2 = (endtime2 - starttime2).microseconds
# print_tf(time1, time2)
print('time1:', a)
print('time2:', b)