[CPU][SNIPPETS] Update the skip markup for Mamul / FC operations

[EgorDuplensky]

taking into account recent changes:

• Matmul also supports bias fusion
• There is no reshape after FC node since corresponding transformation was removed

Also bias fusing should be checked first, since in general bias can be considered as a simple fusing as well.

Main loop also refactored a bit.

[a-sidorova]

Could you please add tests to check that biases aren't fused into Subgraph? We have almost the same test for Convolution

[dmitry-gorokhov]

@EgorDuplensky @a-sidorova @IvanNovoselov This one seems to be waiting for a long time. What are the next steps?

[EgorDuplensky]

@dmitry-gorokhov This change was postponed as not really urgent one. TODO Add unit test TODO Run benchmarking

[akladiev]

This PR will be closed in 2 weeks in case of no activity.

[akladiev]

This PR was closed because it has been stalled for 2 week with no activity.