computeSliceParameters: Fix offset when m(0) != 0

[mgehre-amd]

For affine maps where m(0) != 0,
like affine_map<(d0) -> (d0 + 3) in

  %generic = linalg.generic
    {indexing_maps = [affine_map<(d0) -> (d0 + 3)>,
                      affine_map<(d0) -> (d0)>],
     iterator_types = ["parallel"]} ins(%arg0: tensor<9xf32>) outs(%empty : tensor<6xf32>) {
    ^bb0(%in : f32, %out: f32):
      linalg.yield %in : f32
    } -> tensor<6xf32>

tiling currently computes the wrong slice offsets. When tiling above example with a size of 3, it would compute

scf.for %i = ...
  %slice = tensor.extract_slice %arg0[%i + 3] [6] [1]
  linalg.generic
    {indexing_maps = [affine_map<(d0) -> (d0 + 3)>,
                      affine_map<(d0) -> (d0)>],
     iterator_types = ["parallel"]} ins(%slice: tensor<6xf32>)

and thus apply the +3 twice (once in the extract slice and a second time in the linalg.generic).

This PR fixes this to yield an offset of
tensor.extract_slice %arg0[%i] [6] [1] instead.

[cferry-AMD]

The way I understood this is that it results in a larger slice than necessary. I don't know if we have another choice: since the expressions aren't always affine, but still monotonically increasing, so we can't really afford to change the linalg op's access expression by subtracting the constant there (which would work if the expressions were truly affine). Do we share the same point of view?

[mgehre-amd]

The way I understood this is that it results in a larger slice than necessary. I don't know if we have another choice: since the expressions aren't always affine, but still monotonically increasing, so we can't really afford to change the linalg op's access expression by subtracting the constant there (which would work if the expressions were truly affine). Do we share the same point of view?

Yes, one could have "better" solutions if we were allowed to change the linalg affine_maps.