[autoparallel] added bias comm spec to matmul strategy

[FrankLeeeee]

Problem

The previous PR #1616 considers all matmul sharding strategy generation, however, it failed to consider the the communication action for the bias term.

Solution

This bias term can only have one action which is IDENTITY_FWD_ALLREDUCE_BWD to make sure the gradient is synchronized. The mesh dimensions for all reduce depends on the sharding spec of the output. For example:

1. if the output is two dimensional and has sharding spec RS, then bias also has sharding spec S and thus has no all reduce oepratin.
2. if the output is SS, then the bias requires an allreduce in the backward to sync gradient.

This PR implemented these strategies for the bias term.

Note

The bias term in ops like torch.addmm may not be 1D tensor, I will make the strategy broadcastable in a separate PR.