Compile Tensorflow 1.14 for windows
Required :
1. Download sources from GitHub - tensorflow/tensorflow at r1.14
2. Be sure you have installed Cuda 10.0, and CuDNN7.4.1 for Cuda 10.0 (you have to login for CuDNN download)
        - Note : see Build from source  |  TensorFlow for compatibility table
        - add path below to system path
            - C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\bin
            - C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\libnvvp
        - be sure you have system variables
            CUDA_PATH                     C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0
            CUDA_PATH_V10_0        C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0
3. Download bazel compiler system ver. 0.24.1 as recommended by compatibility table
        - unpack it to separate folder and add this path to the system path
        - Note :  probably it could fail to download some packages and compile time (server is updated continuously, same packages became obsolete), then try to use different version, another versions may do compilation job as well, just try it
4. Additionally you will need to install Cuda_10.1 (temporarily)
        - replace cudaf++.exe from 10.1   (nvcc 10.0 has crash problem)
                - Note : see
        - you may remove Cuda_10.1 after it is done
5. Bazel is using a lot of space on disc C: (Upto ~20G). If you have not enough space on disc C:, consider :
        - move windows temp buffer to another disc (System variables TEMP and TMP)
        - disable pagefile.sys on disc C: and enable it on another disc (for example D:)
        - disable hibernation mode (hiberfil.sys will be removed)
        - clean recycle bin
        - you may also use 'cleanmgr' to empty additional space
        - you may revert back all changes above after compilation is done

Compiling :
1. Configure
        - open CMD window and change to tensorflow main source directory
        - run >> python
            - answer on configuration questions:
                - Usually I'm setting everything by default, but change computing compatibility to GPUs I have (for ex. 3.0, 6.1)
            - you will get ".tf_configure.bazelrc"
                - in .tf_configure.bazelrc
                    - change true->false for compile without XLA    (dont remember why I did it, may be to avoid compilation error)
                        build:xla --define with_xla_support=false
                    - add definitions
                        build --action_env TF_CUDA_VERSION="10.0"
                        build --action_env TF_CUDNN_VERSION="7"
2. Run compilation
        - run >> bazel build --config=opt --config=cuda //tensorflow/tools/pip_package:build_pip_package
        - it will take ~4hours on 4core desktop PC
3. Build wheel package
        - run >> ./bazel-bin/tensorflow/tools/pip_package/build_pip_package ./tensorflow_pkg
Note (just comment) : at configuration stage I made configuration as below (but default values ar also Ok):

c:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\lib\x64\,c:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\include\,c:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\bin\,c:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\

c:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\,c:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\include\