[mlir][vector] Update tests/patterns for vector.transpose (llvm#91359)

Pretty much all logic that we have today for lowering vector.transpose
assumes fixed length vectors (it's done via vector.shuffle that don't
support scalable vectors). This patch updates related tests and patterns
to capture and document this limitation more explicitly.

Note that `vector.transpose` is a valid operation in the context of
scalable vectors, but we are yet to implement the missing lowerings.

Summary of changes:
* `@transpose_nx8x2xf32` is renamed as `@transpose_scalabl`e
  and moved near other tests using `lowering_strategy = "shuffle_1d"
  (to avoid duplicating TD sequences)
* tests specific to X86  (`avx2_lowering_strategy = true`) are moved to
  a dedicated file (to separate generic tests from target-specific
  tests)
* `@transpose10_nx4xnx1xf32` duplicated `@transpose10_4xnx1xf32` and was
  deleted (the latter is renamed as `@transpose10_4x1xf32_scalable` to
  match its fixed-width counterpart: `@transpose10_4x1xf32`)

Author: banach-space

mlir/lib/Dialect/Vector/Transforms/LowerVectorTranspose.cpp

@@ -352,6 +352,7 @@ class TransposeOpLowering : public OpRewritePattern<vector::TransposeOp> {
       return success();
     }
 
+    // TODO: Add support for scalable vectors
     if (inputType.isScalable())
       return failure();