Duplicate fill on contractions

[chelini]

Duplicate fill operations when the use is a contraction and we can fold the fill in the contraction later on in the pipeline using: fold-xsmm-flags. Duplication avoids introducing memref.copies by bufferization. Example,

%0 = tensor.empty()
%1 = linalg.fill ins(...) outs(%0) // fill with zeros.
%2 = linalg.matmul ins(...) outs(%1)
%3 = linalg.matmul ins(...) outs(%1)

Without this PR it bufferizes as:

%0 = memref.alloc()
%1 = memref.alloc()
linalg.fill ins(...) outs(%0) // fill with zeros.
memref.copy %0 into %1
linalg.matmul ins(...) outs(%0)
linalg.matmul ins(...) outs(%1)

With this PR the IR looks like:

// no copies and fills folded as beta = 0.
%0 = memref.alloc()
%1 = memref.alloc()
xsmm.matmul ins(...) outs(%0) // beta = 0
xsmm.matmul ins(...) outs(%1) // beta = 0

The PR has minor performance impact, the only notable improvement is for fp32_mha_tensorflow_seq_len_32. The IR looks cleaner too with 1 less allocation and all the beta flags properly folded. fp32_mha_tensorflow_seq_len_1024 does not improve because dimensionality allows fusion to distribute the fill, see: b1167fe.

This PR is part of #783

[rengolin]

I can't remember what were the examples we discussed about external dependencies to following ops, but we need to make sure that the copy isn't being used by other non contraction/matmul operations, so probably need an artificial test to make sure we check for number of users of each following fill/contraction.