* Allow external mklml outside of prebuilt mkldnn install directory.

* Limit prebuilt mkl-dnn support to Linux.

* update licenses for 2019

* style

* Forward nGraph's C/C++ compiler, build type and generator information to cmake based external projects.

* Fix typo.

* Pass generator related info properly.

* Googletest is using DEBUG_POSTFIX

* TBB uses postfix for debug lib.

* Fix typo.

* update build byproducts to support ninja

* remove unused cmake code

* more cmake cleanup

* display error message if Ninja generator is requested

* fix mkldnn ext project

* revert onnx cmake file

* revert protobuf cmake file

* revert mlsl cmake file

* more fixing

* Update MKLDNN v0.16 -> v0.17

* Updated SHA1 for mac and win

* Updated from 0.17-rc -> 0.17

* Patch to treat OMP as warning instead of error

* Fixed compile error for clang

* Commbined 2 patches into 1 file; renamed file to .patch

* Corrected file permissions

* Move to MKLDNN v0.16. Git hash picks up some bug fixes post v0.16

* Temporary workaround for an MKLDNN bug

* revert MKLDNN ver to 0.14

* reinstate dep

* add EXCLUDE_FROM_ALL to codegen library so we don't need to using conditionals in build. If codegen is needed it will be built.

* update llvm dependencies

* - update MKLDNN version from v0.14 -> v0.15

* - added support to query tensor size from mkldnn desc for MKLDNN Op's
- moved tensor size calculation to tensor_view_layout
- populate mkldnn layout desc witk MKLDNN queried size in the cpu_layout pass

* style fix

* - Fixed failing test cases

(cherry picked from commit d68d3738673c2081f458f9aa458c16361cc47657)

* style fix

* renamed tensor->logical_size() member function to tensor->size()

* Addressed PR comments

* - update mkldnn from v0.15 to v0.16
- update mkl SHA1 hash for lnx, max, win

* style fix

* Addressed Scott's PR comments

* merge branch master with pruthvi/mkldnn_v0.15

* cmake runs for interpreter

* more updates towards building on windows

* remove reference to ngraph core code from codegen. add stand-alone implementations of needed funcions

* fixed potential pointer leak

* clean up file_util

* more file util cleanup, removing unused functions

* interpreter works on mac

* CPU and INTERPRETER build and pass unmit tests on macos

* move get_directory to file_util

* cleanup

* fix mkldnn rpath

* fix compile warning

* close backends when exiting

* set backend output directory of backends to the ngraph output directory

* Aprocter/patch patch (#1119)

* Move more rpath stuff inside if(NOT APPLE)

* fix repatch problem with mkldnn library

* add updated patch command for older versions of cmake

* backend libraries now found in tree

dynamically read header search paths

fix running from install

* finally have something almost acceptable

* Update cmake files to more modern approach

* disable building libraries that are not required

* handle more build cases

* add versions to backend libs. add start of package target.

* add create_backend to backends

* temporary workaround to tbb not linking correctly with gcc

* install codegen lib

* force tbb to link to the cpu backend so that it is available for codegen

* fix clang build error

* fix warning for codegen build

* update cuda header paths

* change error message for opening backend shared library

* set lib path

If user manully provide MKLDNN_INCLUDE_DIR and MKLDNN_LIB_DIR, don't
build mkl-dnn and just add a dummy external project "ext-_mkldnn" to
satisfy target dependency for the rest of the build.

* Enable NNP reverse build, clean up ngraph repo

* clean mkldnn cmake

* Remove usage of CMAKE_MAKE_PROGRAM as it slows down parallel build

* Make make properly propagate to child and add back targeted build.

* Revert "Make make properly propagate to child and add back targeted build."

This reverts commit b4b1d8db0f0d42850e53d4e0f773261c292ccaf6.

* initial support for hybrid transformer

* add broadcast_vector_rowwise_reversed for hybrid test

* headerc

* get function placement fix

* conv ref test generator graph node in labmda fuction

* rename map_parameter_to_source_node

* type change map_parameter_to_source_node

* use interpreter for numerical derivative

* better comments

* Allow caching of external dependencies (everything but TBB, which I can't figure out yet)

* Tweaks to get CPU backend building and linking (but not actually working) on macOS

* Tweaks to codegen/CPU backend to enable macOS (no builtin headers yet)

* Re-enable header caching stuff on macOS

* Suppress unsupported compiler warning for Apple Clang

* Fix includes for macOS compatibility

(Squashing this commit but going to leave the full log here, unlike the way I usually do things)

* Incremental work on compatibility with system requirements

* Lower cmake version requirement so we can work with CentOS 7.4

* Disambiguate isnan call (CentOS's compiler doesn't like it)

* Make external_tbb conditional on NGRAPH_CPU_ENABLE

* Make MKLDNN inclusion conditional on NGRAPH_CPU_ENABLE

* Roll back desperation change to Ubuntu 14.04 binaries

* De-revert copyright date change (arrgh)

* Revert change of TBB's hard-coded compiler choice to gcc

* Cleanup of modified CMake stuff

* Add broken support for building LLVM from source

* Explain modification to json.hpp required compiler version

* Make TBB optional, and on by default. To disable, cmake -DNGRAPH_TBB_ENABLE=FALSE

* Add reasonable default for LLVM's CMAKE_ASM_COMPILER so users don't have to specify it

* Force TBB disabled if CMake is too old to build it (for CentOS)

* Build LLVM OpenMP as external project

* Tweak LLVM build options

* Minor comment edit in cmake scripts

* Workaround (fix?) for codegen crashes on CentOS

* Add some header paths for CentOS to resource_generator; temporary(?) tweak to ignore nonexistent paths

* Remove CMake compatibility figleaf since LLVM will require 3.1+ anyway

* Fix the TBB-enable flag to default to on

* Remove hard-coded choice of clang as tbb compiler

* ln -s -> ln -sf in command to create link for OpenMP source dir

* Make OpenMP build so it doesn't require symlink

* Allow use of prebuilt LLVM if requested at cmake time

* Some INSTALL instructions

* Allow NGRAPH_USE_PREBUILT_LLVM to be set in environment variable

* Tweak CI cmake options to use prebuilt LLVM

* Remove debug messages

* Rework defaulting behavior for LLVM_TARBALL_URL and LLVM_SHA1_HASH

* Remove warning message for unavailable header file directories

Update to newer version of MKLDNN with verbose debug output. Pull from github.com/intel/mkl-dnn instead of github.com/01org/mkl-dnn

* CMake: TBB integration placeholder

* CMake: Integrate TBB

* CMake: Indent

* CMake: Rewrite TBB integration

* CMake: More TBB integration changes

* CMake: Install TBB headers and DSOs

* CMake: Don't install the TBB debug DSO

* CMake: Propagate ngraph's configured compiler setting over to MKL-DNN

* CMake: Restore TBB debug DSO installation

* CMake: Add installed headers to search path.
This needs to be cleaned up along with other header search cleanup

* CPU: Build and execute TBB flowgraphs

* CPU: TBB fixes

* CPU: More TBB fixes

* CPU: Allow both TBB and serial codegen for now

* TBB: get_arguments -> get_input_ops

* CPU: Use node methods

* CPU: Add TBB headers in the build directory to the search path

* TBB: Incorporate various changes from master

* CMake: Indentation fix

* CMake: Indentation fix

* CMake: TBB is mandatory so remove additional predicates

* TBB: Add a test

* CMake: Fix linker flags with GCC

This reverts commit 39383029.

It looks like the commit actually suppressed parallel makes
of MKL-DNN, at least in the case where ngraph itself was being
built with parallel make.  It also introduced problems with
make jobserver warnings.

* Adds CMake variables `MKLDNN_BUILD_COMMAND_EXTRA_FLAGS`
  and `MKLDNN_CMAKE_EXTRA_FLAGS`.