* quick fix to add openmpi as default

* add finalize to distributed class & use unit test

* use intel mlsl github link

* apply style

* address a few comments

* fix test

* update nbench cmake

* remove extras

* fix a bug

* add counter to finalize and cleanup

* test ci error

* address mlsl ci error

* update flag names, as mentioned in pr comment

* revert back the link to mlsl repo and tag

* add flag to finalize

* apply style

* debug with info

* delete when flag is true

* add distributed setup class works, tests pass

* fix style

* remove extra instance

* disable the test due to a bug

* change flag to ompi

* remove the dependency of setting NGRAPH_DISTRIBUTED_ENABLE flag

* cleanup

* change extern to static

* remove the option NGRAPH_DISTRIBUTED_ENABLE setting this flag

* formatting

* update flags not catched by ci

* make unique pointer

* remove unused bool, fix clang error

* first cut at raspberry pi backend

* rename rpi to generic cpu

* disable cursed test

* update licenses for 2019

* style

* Fixed the benchmark file to reflect the updated compile function usage.  Compile returns the compiled function now

* Removed redundant compilation

* files pulled from bob/winbuild

* fix compile problems

* fix a few windows build errors

* add windows file to exclude from git

* add comment why change was made

* revert obsolete change

* more cleanup

* building interpreter and unit test on windows with DLLs

* Add flag for windows to export all symbols. Short term fix.

* enable MD build

* address warnings

* dump all windows build results to a single directory

* fix windows backend dll open issue

* remove debug

* fix file iterator for windows

* fix merge error

* fix test failure

* change header from h to hpp in hopes of making python happy

* address more linux build issues

* fix visibility enable

This reverts commit c5b082c6.

* onnx: return signle nGraph function with multiple outputs
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* onnx: remove excessive code
Signed-off-by: Artur Wojcik <artur.wojcik@intel.com>

* Update ngraph python unit tests

* Fix uninitialized parameter tensor data in nbench

* Remove const modifier in the loop to fix compilation error

* Add condition check before writing tensor data

* Remove the condition check when initializing pramater tensor data

Use @loader_path for the calling library/executable
Use @rpath for dependent libraries.

* Add PlaidML backend

* CR comments

Used m_ prefix for members; removed trailing underscores
Updated license headers
Moved associated header inclusions to project blocks
Wrapped comments to 100 chars
Added missing newlines between functions
Removed nested namespaces in operation implementations

* Add earhart to CODEOWNERS

* Rebase updates

* style

Fixes an error when building on macOS (with CPU backend disabled, but not sure if that makes a difference):

```
/Users/amprocte/Work/ngraph/src/tools/nbench/benchmark.cpp:36:29: error: use of undeclared identifier '_MM_FLUSH_ZERO_ON'
    _MM_SET_FLUSH_ZERO_MODE(_MM_FLUSH_ZERO_ON);
                            ^
/Users/amprocte/Work/ngraph/src/tools/nbench/benchmark.cpp:37:33: error: use of undeclared identifier '_MM_DENORMALS_ZERO_ON'
    _MM_SET_DENORMALS_ZERO_MODE(_MM_DENORMALS_ZERO_ON);
```

* IntelGPU backend: Enable nBench statistic

* PR1848. Combine counters type time

* [ONNX] Tool to serialize ONNX models

* Review fix.

* Add back dependencies to onnx_import_interface target.

* Review fix pt. 2

* Check also for errors when opening file stream.

* rename files

* rename runtime TensorView to Tensor

* rename HostTensorView to HostTensor

* remove get_tensor method

* remove get_element_count as it is redundant to get_size

* fix a few issues with examples and python wrapper

* rename get_size to get_element_count and add get_size_in_bytes method

* update nbench output to be more consistent for single vs. directory benchmark.

* wip

* add types used in model

* fix test

* style

* catch all exceptions when running any option

* add unsupported_op exception

* unsupported_op test

* add printout of unsupported op in model

* fix GPU dispatcher check

* fix test designation

* catch exceptions on single file runs too

* add unsupported_op exception where needed

* remove unsupported_op class

* add unassigned op exception

* add unit test

* catch unsupported op in nbench

* add cpu test back

* update all latest merges

* mode change

* add option to copy intput/output data for each iteration

* add support for stale buffers

* only print details if details enabled

* refactor print routines to main file

* refactor for multi dir support

* dir support prints nice results

* use line comments instead of multiline comments for license header

* update more

* update new files

* more header updates

* style

This reverts commit 5ab5a129.

* static link cpu library to ngraph

* remove debug

* link ngraph and cpu backend into a single shared object

* add -fPIC and whole-archive for CPU backend

* Added conditional for --whole-archive for Mac OS.

* Added more conditonal for MacOS.

* fix linking problem and unit test failures caused by multiple copies of the same function in CPU backend and INTERPRETER

* fix nbench build

* add nbench to unit test build

* add version number to libngraph

* Optionally get rid of codegen from the CPU backend

* Rename option variable

* Merge fixes

* Merge

* Remove extra changes

* remove dex only exclusions (#1429)

* Unconditionally pick  m_direct_execution if NGRAPH_DEX_ONLY

* Style fix

* Add an option to exclude the first iteration

* Switch to warmup iterations

* Cleanup

* add option to run all models in a directory

* add print for exception from benchmark

* compiles but does not link

* using async gpu timers

* remove sync for cuda calls, add async gpu stopwatch, add count to timing-detail

* add debug sync

* make timer static

* move timer to runtime context

* nbench cleanup

* update style