[llvm] [SelectionDAG][x86] Ensure vector reduction optimization (PR #144231)

[Suhajda Tamás via llvm-commits]

sutajo wrote:

I tried to implement custom lowering of vector reduction ops for x86, but it turned out to be more complicated than I expected. It is difficult to know in advance which reductions can be efficiently expanded.

I don't have the time currently to look more into this, but for now this PR fixes the original issue.
This fix can be removed once the reductions have custom lowering.

https://github.com/llvm/llvm-project/pull/144231