CODEWITHSUNDEEP
pythontensorflowreproducible-research
Duong Nguyen
Duong Nguyen

Reputation: 11

Different results on different machines tensorflow

I'm running a Tensorflow 1.12 code (I'm not using GPU). I have set

import os
os.environ['TF_DETERMINISTIC_OPS'] = '1'
os.environ['TF_CUDNN_DETERMINISTIC'] = '1'
os.environ['PYTHONHASHSEED'] = '42'

import tensorflow as tf
tf.reset_default_graph()
session_conf = tf.ConfigProto(
    intra_op_parallelism_threads=1, 
    inter_op_parallelism_threads=1)
np.random.seed(42)
tf.set_random_seed(42)

<<my code>>

If I run the code on the same hardware, the result is identical. However, if I run the code on different hardware (e.g. Xeon E5 vs Epyc Rome 7402), I git different results.

Could you please help me solve this problem? Thank you very much in advance!

Upvotes: 1

Views: 560

Answers (1)

Kholdarbekov
Kholdarbekov

Reputation: 1150

Tensorflow performance is a little dependent on hardware configuration also.

The performance of TensorFlow Serving is highly dependent on the application it runs, the environment in which it is deployed and other software with which it shares access to the underlying hardware resources.

Read more from: https://www.tensorflow.org/tfx/serving/performance

Upvotes: 2

Related Questions