[llvm] [GlobalISel] Add multi-way splitting support for wide scalar shifts. (PR #155353)

[Amara Emerson via llvm-commits]

aemerson wrote:

> That is likely why it is so slow for large shifts, if I have that correct.

To clarify, this patch is a large compile time improvement for large shifts vs today. At the moment for shifts in the range of 4k bits it's much, much worse.

https://github.com/llvm/llvm-project/pull/155353