CODEWITHSUNDEEP
pythontensorflowkeras
sten
sten

Reputation: 7486

Why Keras model on "bare" CPU is faster?

Running a Keras model... the bad thing is that it is faster not to use CPU extentions (it should be the other way around). Look at the output below.

Is there a config file where I can set inter_op_parallelism option ?

 Using TensorFlow backend.
 2018-10-18 17:21:32.620461: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
 2018-10-18 17:21:32.621535: I tensorflow/core/common_runtime/process_util.cc:69] Creating new thread pool with default inter op setting: 2. Tune using inter_op_parallelism_threads for best performance.
 Results: -33.20 (23.69) MSE

 real   2m55.990s
 user   4m8.784s
 sys    3m50.192s

 Using TensorFlow backend.
 2018-10-18 17:25:04.773578: I tensorflow/core/platform/cpu_feature_guard.cc:140] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
 Results: -32.57 (23.16) MSE

 real   1m48.847s
 user   2m15.792s
 sys    0m13.440s

Upvotes: 1

Views: 1271

Answers (1)

Charles Nehez
Charles Nehez

Reputation: 63

Here you are the code I'm using with keras, just put it on the top of your code. 

from keras import backend as K
import tensorflow as tf

NUM_PARALLEL_EXEC_UNITS = 6

config = tf.ConfigProto(intra_op_parallelism_threads = NUM_PARALLEL_EXEC_UNITS, 
         inter_op_parallelism_threads = 1, 
         allow_soft_placement = True, 
         device_count = {'CPU': NUM_PARALLEL_EXEC_UNITS })

session = tf.Session(config=config)

K.set_session(session)

import os

os.environ["OMP_NUM_THREADS"] = str(NUM_PARALLEL_EXEC_UNITS)
os.environ["KMP_BLOCKTIME"] = "30"
os.environ["KMP_SETTINGS"] = "1"
os.environ["KMP_AFFINITY"]= "granularity=fine,verbose,compact,1,0"

Note: I'm a little bit disappointed with the results. I could reach maximum 150% speeding up only playing with these parameters.

Upvotes: 1

Related Questions