Maksim Kutakov

Is it possible to cover the change with an SL/Unit test, e.g. use the pattern from the issue ticket?

build_jenkins

I second @usstq in his argument about inability of the ngraph representation to distinguish which data movement ops are optimized (are not run in the end) and which are not....

@EgorDuplensky , could you please review?

> ReLuConcatConvSumInPlaceTest Yes, sure, but please use **nspc** layout instead as the main one for optimized convolution implementations.