[SYCL-MLIR]: Improve Polygeist LICM to handle invariant operations in scf::ForOp and AffineForOp

[etiotto]

This PR improves the ParallelLICM pass in polygeist as follows:

• adds support for non-parallel loop constructs scf::ForOp and improves handleing of AffineForOp
• fixes a problem, in sequential loops function calls are conservatively considered to modify memory
• add more tests for both reduction loops and loops that do not yield a value
• renames ParallelLICM pass to simply LICM

[victor-eds]

Nice changes to this pass. I think we're missing something else regarding operations side-effects: speculability. We should address this in a later PR, but it'd make sense to check for this also in this pass.

[etiotto]

The failure:

---

Failed Tests (1):
SYCL :: BFloat16/bfloat16_conversions.cpp

is OK because another PR has a fix for this failure and that PR is not yet merged.

The x64 SYCL LLVM test failures are caused by a machine problem:

terminate called after throwing an instance of 'sycl::_V1::runtime_error'
  what():  No device of requested type available. -1 (PI_ERROR_DEVICE_NOT_FOUND)

I am merging the PR.