Merge branch 'master' into sarkars/unify_pooling_translate_functions

Author: sayantan-nervana

Diff for: CMakeLists.txt

@@ -239,7 +239,7 @@ if (NOT USE_PRE_BUILT_NGRAPH)
     ExternalProject_Add(
         ext_ngraph
         GIT_REPOSITORY https://github.com/NervanaSystems/ngraph
-        GIT_TAG v0.15.1-rc.2
+        GIT_TAG v0.16.0-rc.0
         CMAKE_ARGS 
             -DNGRAPH_DISTRIBUTED_ENABLE=${NGRAPH_DISTRIBUTED_ENABLE}
             -DNGRAPH_INSTALL_PREFIX=${NGRAPH_ARTIFACTS_DIR}

Diff for: README.md

@@ -27,7 +27,7 @@ a variety of nGraph-enabled backends: CPU, GPU, and custom silicon like the
         virtualenv --system-site-packages -p /usr/bin/python2 your_virtualenv  
         source your_virtualenv/bin/activate # bash, sh, ksh, or zsh
 
-2. Install TensorFlow v1.12.0:
+2. Install TensorFlow v1.13.1:
 
         pip install -U tensorflow
 
@@ -41,10 +41,10 @@ a variety of nGraph-enabled backends: CPU, GPU, and custom silicon like the
 
 This will produce something like this:
 
-        TensorFlow version:  1.12.0
-        nGraph bridge version: b'0.11.0'
-        nGraph version used for this build: b'0.14.0+56a54ca'
-        TensorFlow version used for this build: v1.12.0-0-ga6d8ffae09
+        TensorFlow version:  1.13.1
+        nGraph bridge version: b'0.12.0-rc1'
+        nGraph version used for this build: b'0.21.0-rc.0+b638705'
+        TensorFlow version used for this build: v1.13.1-0-g6612da8951
 
 Next you can try out the TensorFlow models by adding one line to your existing 
 TensorFlow model scripts and running them the usual way:
@@ -63,11 +63,11 @@ bridge using the TensorFlow source tree as follows:
 The installation prerequisites are the same as described in the TensorFlow 
 [prepare environment] for linux.
 
-1. TensorFlow uses a build system called "bazel". These instructions were tested with [bazel version 0.16.0]. 
+1. TensorFlow uses a build system called "bazel". These instructions were tested with [bazel version 0.21.0]. 
 
-        wget https://github.com/bazelbuild/bazel/releases/download/0.16.0/bazel-0.16.0-installer-linux-x86_64.sh      
-        chmod +x bazel-0.16.0-installer-linux-x86_64.sh
-        ./bazel-0.16.0-installer-linux-x86_64.sh --user
+        wget https://github.com/bazelbuild/bazel/releases/download/0.21.0/bazel-0.21.0-installer-linux-x86_64.sh      
+        chmod +x bazel-0.21.0-installer-linux-x86_64.sh
+        ./bazel-0.21.0-installer-linux-x86_64.sh --user
 
 2. Add and source the ``bin`` path to your ``~/.bashrc`` file in order to be 
    able to call bazel from the user's installation we set up:
@@ -84,7 +84,7 @@ The installation prerequisites are the same as described in the TensorFlow
 
         git clone https://github.com/NervanaSystems/ngraph-tf.git
         cd ngraph-tf
-        git checkout v0.11.0
+        git checkout v0.12.0-rc1
 
 
 2. Next run the following Python script to build TensorFlow, nGraph and the bridge. Please use Python 3.5:
@@ -115,73 +115,9 @@ with nGraph backends.
 
 Please add the following line to enable nGraph: `import ngraph_bridge`
 
-## Option 3: Using the upstreamed version
-
-nGraph is updated in the TensorFlow source tree using pull requests periodically. 
-
-In order to build that version of nGraph, follow the steps below, which involves building TensorFlow from source with certain settings.
-
-1. Install bazel using the same instructions outlined in option 2, step 1 above.
-
-2. Create a virtual environment using the instructions outlined in option 1, step 1 above.
-
-3. Get tensorflow v1.12.0
-
-        git clone https://github.com/tensorflow/tensorflow.git
-        cd tensorflow
-        git checkout v1.12.0
-
-Note: To get the latest version of nGraph, use the tip of `master` branch of TensorFlow. The exact version of `bazel` changes for a specific version of TensorFlow. Please consult the build instructions from TensorFlow web site for specific bazel requirements.
-
-4. Now run `./configure` and choose `no` for the following when prompted to build TensorFlow.
-
-    XLA support:
-
-        Do you wish to build TensorFlow with XLA JIT support? [Y/n]: n
-        No XLA JIT support will be enabled for TensorFlow.
-
-    CUDA support:
-    
-        Do you wish to build TensorFlow with CUDA support? [y/N]: N
-        No CUDA support will be enabled for TensorFlow.
-    
-    :warning: Note that if you are running TensorFlow on a Skylake family processor then select
-    `-march=broadwell` when prompted to specify the optimization flags:
-    
-        Please specify optimization flags to use during compilation 
-        when bazel option "--config=opt" is specified 
-        [Default is -march=native]: -march=broadwell
-    
-    This is due to an issue in TensorFlow tracked here: 
-    https://github.com/tensorflow/tensorflow/issues/17273
-
-5. Prepare the pip package
-
-        bazel build --config=opt --config=ngraph //tensorflow/tools/pip_package:build_pip_package 
-        bazel-bin/tensorflow/tools/pip_package/build_pip_package ./
-
-Note: The specific questions for the `configure` step and the build command mentioned above changes for different versions of TensorFlow. 
-
-6. Once the pip package is built, install using
-
-        pip install -U ./tensorflow-1.*whl
-
-For this final option, there is **no need to separately build `ngraph-tf` or to 
-use `pip` to install the nGraph module**. With this configuration, your TensorFlow model scripts will work without any changes, ie, you do not need to add `import ngraph_bridge`, like option 1 and 2. 
-
-Note: The version that is available in the upstreamed version of TensorFlow usually
-lags the features and bug fixes available in the `master` branch of this repository.
-
-You can run a few of your own DL models to validate the end-to-end 
-functionality. Also, you can use the `ngraph-tf/examples` directory and try to 
-run the following model: 
-
-        cd examples
-        python3 keras_sample.py 
-
 ## Using OS X 
 
-The build and installation instructions are idential for Ubuntu 16.04 and OS X. However, please
+The build and installation instructions are identical for Ubuntu 16.04 and OS X. However, please
 note that the Python setup is not always the same across various Mac OS versions. TensorFlow build
 instructions recommend using Homebrew and often people use Pyenv. There is also Anaconda/Miniconda 
 which some users prefer. Ensure that you can build TenorFlow successfully on OS X with a suitable 
@@ -223,7 +159,7 @@ See the full documentation here:  <http://ngraph.nervanasys.com/docs/latest>
 [DSO]:http://csweb.cs.wfu.edu/~torgerse/Kokua/More_SGI/007-2360-010/sgi_html/ch03.html
 [Github issues]: https://github.com/NervanaSystems/ngraph-tf/issues
 [pull request]: https://github.com/NervanaSystems/ngraph-tf/pulls
-[bazel version 0.16.0]: https://github.com/bazelbuild/bazel/releases/tag/0.16.0
+[bazel version 0.21.0]: https://github.com/bazelbuild/bazel/releases/tag/0.21.0
 [prepare environment]: https://www.tensorflow.org/install/install_sources#prepare_environment_for_linux
 [diagnostics]:diagnostics/README.md
 [ops]:http://ngraph.nervanasys.com/docs/latest/ops/index.html

Diff for: build_ngtf.py

@@ -44,7 +44,7 @@ def command_executor(cmd, verbose=False, msg=None, stdout=None):
         raise Exception("Error running command: " + cmd)
 
 
-def build_ngraph(src_location, cmake_flags, verbose):
+def build_ngraph(build_dir, src_location, cmake_flags, verbose):
     pwd = os.getcwd()
 
     src_location = os.path.abspath(src_location)
@@ -53,15 +53,15 @@ def build_ngraph(src_location, cmake_flags, verbose):
     os.chdir(src_location)
 
     # mkdir build directory
-    path = 'build'
+    path =  build_dir
     try:
         os.makedirs(path)
     except OSError as exc:  # Python >2.5
         if exc.errno == errno.EEXIST and os.path.isdir(path):
             pass
 
     # Run cmake
-    os.chdir('build')
+    os.chdir(build_dir)
 
     cmake_cmd = ["cmake"]
     cmake_cmd.extend(cmake_flags)
@@ -278,7 +278,7 @@ def install_tensorflow(venv_dir, artifacts_dir):
     return str(cxx_abi)
 
 
-def build_ngraph_tf(artifacts_location, ngtf_src_loc, venv_dir, cmake_flags,
+def build_ngraph_tf(build_dir, artifacts_location, ngtf_src_loc, venv_dir, cmake_flags,
                     verbose):
     pwd = os.getcwd()
 
@@ -296,15 +296,15 @@ def build_ngraph_tf(artifacts_location, ngtf_src_loc, venv_dir, cmake_flags,
     os.chdir(ngtf_src_loc)
 
     # mkdir build directory
-    path = 'build'
+    path =  build_dir
     try:
         os.makedirs(path)
     except OSError as exc:  # Python >2.5
         if exc.errno == errno.EEXIST and os.path.isdir(path):
             pass
 
     # Run cmake
-    os.chdir('build')
+    os.chdir(path)
     cmake_cmd = ["cmake"]
     cmake_cmd.extend(cmake_flags)
     cmake_cmd.extend([ngtf_src_loc])
@@ -408,8 +408,8 @@ def main():
     parser.add_argument(
         '--use_prebuilt_binaries',
         help=
-        "Skip building nGraph and TensorFlow. Rather use \"build\" directory.\n"
-        + "The following directory structure is assumed:\n" + "build\n" +
+        "Skip building nGraph and TensorFlow. Rather use \"build_cmake\" directory.\n"
+        + "The following directory structure is assumed:\n" + "build_cmake\n" +
         "  |\n" + "   -- artifacts\n" + "  |   |\n" +
         "  |   |-- bin (contains binaries from nGraph build)\n" +
         "  |   |-- include (contains include files from nGraph build)\n" +
@@ -440,11 +440,11 @@ def main():
     #-------------------------------
 
     # Component versions
-    ngraph_version = "v0.15.1-rc.2"
-    tf_version = "v1.12.0"
+    ngraph_version = "v0.16.0-rc.0"
+    tf_version = "v1.13.1"
 
     # Default directories
-    build_dir = 'build'
+    build_dir = 'build_cmake'
 
     # Override the pre-built location is specified
     use_prebuilt_binaries = False
@@ -536,7 +536,7 @@ def main():
         else:
             ngraph_cmake_flags.extend(["-DNGRAPH_GPU_ENABLE=NO"])
 
-        build_ngraph("./ngraph", ngraph_cmake_flags, verbosity)
+        build_ngraph(build_dir, "./ngraph", ngraph_cmake_flags, verbosity)
 
     # Next build CMAKE options for the bridge
     tf_src_dir = os.path.abspath("tensorflow")
@@ -559,7 +559,7 @@ def main():
         ngraph_tf_cmake_flags.extend(["-DNGRAPH_DISTRIBUTED_ENABLE=FALSE"])
 
     # Now build the bridge
-    ng_tf_whl = build_ngraph_tf(artifacts_location, "../", venv_dir,
+    ng_tf_whl = build_ngraph_tf(build_dir, artifacts_location, "../", venv_dir,
                                 ngraph_tf_cmake_flags, verbosity)
 
     print("SUCCESSFULLY generated wheel: %s" % ng_tf_whl)

Diff for: python/setup.in.py

@@ -39,7 +39,7 @@ def get_tag(self):
 
 setup( 
     name='ngraph_tensorflow_bridge',
-    version='0.12.0-rc0',
+    version='0.12.0rc1',
     description='Intel nGraph compiler and runtime for TensorFlow',
     long_description=long_description,
     long_description_content_type="text/markdown",

Diff for: src/CMakeLists.txt

@@ -46,6 +46,7 @@ set(SRC
    ngraph_rewrite_pass.cc
    ngraph_tracked_variable.cc
    ngraph_utils.cc
+   ngraph_timer.cc
    tf_graphcycles.cc
    tf_deadness_analysis.cc
    version.cc