Update to newer version of MKLDNN with verbose debug output. Pull fro…

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

* Support for MKLDNN convolution fprop in CPU backend based on Jai's implementation (no layout yet)

* 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

* Remove TupleType, ValueType

* Fix compile error.

* Simplify Dot Autodiff

* remove commented code

* fix convlution reference script

* convolution backprop

* cleanup

* fix build warnings

* Missing include

* fix build warning part 2

* move numeric_compare to its own header
code review feedback

* fix build warnings 3

* fix build warnings 4

* clang-format

* cast to avoid implicit cast warning

* adding work and organization for docs

* Summary of changes

- removed rst file from css folder
- added substitutions to conf file so folks don't have to type out the whole name
- added branding notice so link in footer works
- added glossary so we can have a place for terms relative to libngraph
- added placeholder for MXNet and TF frontend under new section
- cleaned up api page

* Fix code fencing error and indentation issue causing build to throw error

* adding MXNet integration guide details and other miscellaneous structure

* fix unicode error in rst epilog

* testing with unicode marker for rst_epilog string

* replace api scaffolding with TODO to avoid breaking Jenkins

* wip

* using namespace cleanup

* add base ngraphcpp docs theme

* add base ngraphcpp docs theme and doc contributor README

* Add proper chars for tm and r symbols in main project sidebar

* nevermind that last commit

* added separate sphinx_doc and doxygen_doc targets to the contrib/docker/Makefile to build html docs for PR #377

* explicitly marked non-ascii string as unicode for sphinx conf.py

* wip

* wip

* remove get_vector from runtime::TensorView class as it was for unit test only

* cleanup

* move writting vector to runtime::TensorView to the unit test dir

* merge fix

* PR review change

* update from PR comment

* update changes file

* cleanup in-memory header files

* add switch to enable memory sanitizer (works like valgrind)

* removed header file cleanup as it was causing a segfault on program termination

* Add mention of blob ref of original file from caffe2

* Mention location of source listing originally from LLVM project

* Average pool type checking and kernel; type checking tests

* Fix and enable average-pool tests

* Docstring fix

* Extend AvgPool op type checking to support padding

* Untested code for padded avg-pool

* Unit tests for padded avg-pool

* Add CPU implementation

* Temp delete

* Docstring fix

* Docstring fix

* Add tests mixing padding and stride

* Temporary cut to ease merge

* Restore temporary cut for merge

* Empty commit to try to force CI to wake up

* remove refs on types in tensor,tensor_view_type,convert

* fix build breaks

* change default is_functionally_identical to return false so if an op forgets to override it gets a behavior that might be slower to compile but it will at least work

* add test for empty tuple

* fix null function breaking

* remove a node from users

* style

* add mxnet seq2seq forward and backward

* add benchmarks for seq2seq forward and backward

* Numerically stable sum so we can pass mxnet unit tests

* Add a small initial residual