tensorflow: Fail in configuration due to different CUDA libraries path

System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow): No
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Ubuntu 17.04
• TensorFlow installed from (source or binary): source
• TensorFlow version (use command below): 49f9c6f890c938955fa2d448ac5b556b9a6d9aa0
• Python version: Python 3.5.3
• Bazel version (if compiling from source): bazel release 0.6.1
• CUDA/cuDNN version: CUDA 8, cuDNN 7
• GPU model and memory: GeForce GTX 1080

Describe the problem

As for now, TF assumes that CUDA liibs are located at CUDA_PATH/lib64/, however Ubuntu installs CUDA to /usr/lib/x86_64-linux-gnu/, which makes configuration impossible: I can’t specify cuda path such that it’ll find /usr/lib/x86_64-linux-gnu/libcudart.so.8.0