* Fix warnings in zero-sized tests; fix zero-sized tests return vector type for comparison ops

* Revert changes to cpu_emitter

* Enable CPU backprop tests

* Fix to dot codegen for cases where n_reduction_axes != 1

Dynamically create and destroy mkldnn engine objects to work around f…

CPU: Introduce reshape fallback before we switch to Eigen tensors

* Fix project README to be current, new doxyfile config test, cleanup static dir of theme

* make sure new README gets added

* fixes on doc cleanup branch

* Add feedback from github reviews

* Fix core README again

* test relu fusion for argon backend and enable permutation over max op

1. AvgPoolBackprop type checking
2. Padding for max-pool op and kernel
3. Max pool backprop kernel, and integrating everything with CPU backend/serializer
3a. Add optional backpointer from the max pool backprop op to the forward prop op

* modify existing autodiff unit tests to test fprop cache

* cleanup

* fix compile error introduced with bad merge

* remove invalid negative/negative backwards power test

(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

Merge branch 'jbobba/mkldnn-outlining' of https://github.com/NervanaSystems/private-ngraph-cpp into jbobba/mkldnn-outlining

inline

Inliner pass + tests

debugging

fix inliner failures due to the fact a random function is picked as an outermost one

copyright headers

fix clone of function with multiple outputs

add robust replacement for is_functionally_identical that relies on comparing emitting functions as string (#441)

* mark convolution and pooling as non-functionally equivalent
temporary workaround for MKLDNN issue

* fix build warning

* More graph construction

* Review comments

* fix resource generator dependencies

* add cr to end of file

* simplification pass

* serializer change to test models

* some small test fixes

* addressing Scott's feedback

* missed one nn

* formatting fixes

* simplification -> reshape_elimination

* WIP on finding a good format for op docs in RST

* A few more scribbles

* fix up branch for Amazon code share

* add conf.py configuration details from aproctor's branch for doxy-breathe integration

* update section on how to build the documentation with breathe install details

* Remove empty file on training, update framework integration notes

* Add CentOS stub, fix spelling, core op definition, add to glossary.

* more documentation cleanup on README and installation and testing

* more cleanup of docs for TernsorFlow

* Simplify Dot Autodiff (#412)

* Simplify Dot Autodiff

* remove commented code

* Remove TupleType, ValueType (#411)

* Remove TupleType, ValueType

* Fix compile error.

* Change convolution reference to work with f32 (#409)

* Drwebb/gpu backend dot op (#413)

* Drwebb/gpu backend dot op (#387)

* GPU Dot prod emitter switch statement

* cuBLAS dot kernel call

* Flush out arg substitution into gpu dot kernel call

* Drwebb/gpu backend dot op (#392)

* Take in CodeWriter into gpu op emitters

* Introduce GPU function gen based on pass functions

* Additional gpu emitter stubs

* link cublas in to unit test and ngraph

* Use static code gen methods for GPU, add new GPU op stubs

* use pass manager to declare functions / cublas Updates

* Prune down gpu_external_function wip

* Switch back to GPU tensor views in GPU backend

* Pass in cublas handle to GPU external function

* cuMalloc memory in gpu tensor view

* Use cuda runtime malloc and free for tensor view managment c

* change GPU tensor view init, and use GPU tensor view for GPU call frame

* include headers as system dirs

* GPU tensor printing utility function

* cublasSetPointer to device mode / Fix copyright notification lowercasing

* Passing GPU dot product test using cuBLAS

Clean up

* Changes from review

* Add an overivew.

* Intro for building graphs.

* Refactor docs so that Doxygen and Sphinx are integrated (Sphinx depends on Doxygen with the docstrings stuff)

Still need to resolve a lingering assumption that the build dir is contained in private-ngraph-cpp. It's proving to be surprisingly tricky.

* Added the TensorFlow XLA build information and example of how to run MNIST MLP with TF/nGraph

* Updated TF integration guide for clarity. Added files from cyphers-amazon branch. Add minor changes to sphinx-doxy to test apis

* Small revision of overview and add graphic from arXiv paper

* WIP more editing, picking up from where I left off last week

* Fix garbled sentence edit

* WIP Edit for readability and such
:

* Better font rendering on all architectures included with our custom theme

* Cleanup current version of documentation.  Add NeoSans font binaries to make local font rendering of h1 h2 etc

* Missed merge conflict

* Add something on functions, don't forward-reference parameters

* What we have so far into a PR for review

* Need file for cmake

* Missing header

* Remove duplicate file

* added breathe package to contrib/docker/Dockerfile.ngraph_cpp

* bprop for avg pool

remove debug statements + formatting

* fix CPU test failures

* numeric tests

* use make_shared; unprotect c-tor