Why my TensorFlow got warning for first run but then works?

Question

Excuse for the stupid question from a newbie. I just got my TensorFlow, both GPU and CPU, installed on Ubuntu 20.04 LTS on WSL2 with CUDA. My GPU is GeForce 940MX. When I try to test the TensorFlow installation, the first attempt got some warning as the message below. However, the exact same syntax following went through smoothly. I wonder what happened here? And is it using CPU or GPU actually?

>>> import tensorflow as tf
2021-01-22 12:01:14.284464: I tensorflow/stream_executor/platform/default/dso_loader.cc:49] Successfully opened dynamic library libcudart.so.11.0
>>> print(tf.reduce_sum(tf.random.normal([1000, 1000])))
2021-01-22 12:01:20.034653: I tensorflow/compiler/jit/xla_cpu_device.cc:41] Not creating XLA devices, tf_xla_enable_xla_devices not set
2021-01-22 12:01:20.035316: W tensorflow/stream_executor/platform/default/dso_loader.cc:60] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/cuda-11.2/lib64
2021-01-22 12:01:20.035400: W tensorflow/stream_executor/cuda/cuda_driver.cc:326] failed call to cuInit: UNKNOWN ERROR (303)
2021-01-22 12:01:20.035518: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:156] kernel driver does not appear to be running on this host (DESKTOP-3E6PSHT): /proc/driver/nvidia/version does not exist
2021-01-22 12:01:20.035984: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
2021-01-22 12:01:20.036526: I tensorflow/compiler/jit/xla_gpu_device.cc:99] Not creating XLA devices, tf_xla_enable_xla_devices not set
tf.Tensor(-872.4863, shape=(), dtype=float32)
>>> print(tf.reduce_sum(tf.random.normal([1000, 1000])))
tf.Tensor(590.53516, shape=(), dtype=float32)
>>> print(tf.reduce_sum(tf.random.normal([1000, 1000])))
tf.Tensor(294.59973, shape=(), dtype=float32)
>>> print(tf.reduce_sum(tf.random.normal([1000, 1000])))
tf.Tensor(261.34412, shape=(), dtype=float32)

Anything helps!

question from:https://stackoverflow.com/questions/65851043/why-my-tensorflow-got-warning-for-first-run-but-then-works

Answer 1

Waitting for answers