* Added build option to disable fast math

* Change option name

* Added float16 support for constant op

* Fix Compilation issue.

* Style.

* Change serializer disable so that it is enabled by default if you do nothing

* fix serializer disable

* fix disabled build

* layout-aware cpu tensor copy_from

* - add unit test case for tensor "copy_from"
- set default layout after copy_from src to destination

* - added test case to copy between two different layouts

* - fix clang error
 - added test case to check copy_from if the two tensor layouts are different
- fix the LayoutDesc comparision

* fix windows build error

* - move "cop_from" unit tests from tensor to cpu_test

* fix compilation error

* Added more unit tests

* fix read and write API in cpu_tensor_view

* Failing conv test case

* Opt for mkldnn::algorithm::convolution_direct if input channel is less than 8

* Comment

* Unified interpolate layer op

* Minor attr name changes

* Use Eigen kernel for Gather when axis is 0.

* Fix style error.

* Rename variables.

Fix bugs.

Use helper function.

* Add one unit test.

* Check if openmp is defined.

* Change unit tests.

* Use Eigen kernel for uint8_t type.

Add one uint8_t unit test.

Address PR feedback.

* Update cpu emitter.

* Use updated reduced-size clDNN

* Remove the obsolete BOOST_VERSION variable

* remove tensor offset from tensor read/write calls because it was never used

* fix build errors

* fix build errors

* fix python test errors

* more python fixes

* revert change

* Make old version of read/write deprecated

* fix python read overload

* one more try to fix python binding

* fix python

* yet another try

* why is this so hard

* fix?

* add text to changes.md

* add BatchMatMul unit test to all builds

* Check that interpreter is available

* fix error

* enable tests for features not yet implemented

* wip

* build works with serialize disabled

* one more try

* disable unit test when json disabled

* handle case where serialize is disabled in python test

* one more fix

* Fix klocwork issues

* Need dynamic cast

* Update to use new backend constructor pattern

* Update PlaidML unit test exclusions

*  Use Eigen kernel for ScatterAdd.

* Emit ScatterAdd Eigen kernel for CODEGEN.

Add comments.

Address PR feedback.

* Add more unit tests.

Fix style error.

Add ScatterAdd to AssignOpMap.

* Combine non-scalar cases together.

* Address PR feedback.

* Fix a bug.

* Use reshape to make the shapes of two slices match.

* Rename variables.

Fix bugs.

Use helper function.

Add one unit test.

* Add reshape back.

* Added getters for dynslice

* Style check fixes

* Combine transpose and reshape pattern into a single reshape

* optimize reshapes only if in/out shapes dont match

* Default to svg format for visualizing graphs and provide an env variable to change it

* Handle 0 and -1 in shape inputs for dynamic reshape op

* Flag to control semantics of zero values in dynreshape op

Unused variables.

[MLIR] Improve MLIR dump

It refactor MLIR dump code in MLIRCompiler and adds missing dumps for
some steps of the lowering. It also removes some redundant dump for
affine dialect.

* Support for no-arg op construction

* Add a-ops.

* Two const init variations to try to solve windows link error

* Try export for windows

* Create outputs as needed, test

* Remove attribute helpers

* another windows try

* windows

* windows

* Partial node cleanup

* Review comments
Node destruction that doesn't blow out the stack on deep graphs

* Review comments

* Update src/ngraph/node.hpp
Co-Authored-By: Adam Procter <adam.m.procter@intel.com>

* Have huge test check that nodes were deleted

* Switch constructors to use Output instead of Node

* Review comments

* Add some comments

This patch introduces tablegen definitions for element-wise reduction
operations (only definitions, not building/lowering code). This includes
argmin, argmax, min, max, sum, prod, all and any.

From the design point of view, a single base class seems to be enough to
cover all the common ground of these operations. A different base class will
be necessary for avg pool reduction operations.

* Improve verification for unary, binary, cmp and select ops. Refactor .td file to enable that.

MLIR Compile only supported ops