[Pipelines] Introduce SROA after (final, run-time) loop unrolling
Now that we are done with loop unrolling, be it either by LoopVectorizer,
or LoopUnroll passes, some variable-offset GEP's into alloca's could have
become constant-offset, thus enabling SROA and alloca promotion,
yet we don't capitalize on that, which is surprizing.
While it would be good to not introduce one more SROA invocation,
but instead move the one from `PassBuilder::buildFunctionSimplificationPipeline()`,
the existing test coverage says that is a bad idea,
though it would be fine compile-time wise: https://llvm-compile-time-tracker.com/compare.php?from=b150d34c47efbd8fa09604bce805c0920360f8d7&to=5a9a5c855158b482552be8c7af3e73d67fa44805&stat=instructions
So instead, i add yet another SROA run.
I have checked, and it needs to be at least after said final loop unrolling.
This is still fine compile-time wise: https://llvm-compile-time-tracker.com/compare.php?from=70324cd88328c0924e605fa81b696572560aa5c9&to=fb489bbef687ad821c3173a931709f9cad9aee8a&stat=instructions
I've encountered this in a real code, `SROA-after-final-loop-unrolling.ll` has been reduced from https://godbolt.org/z/fsdMhETh3
Reviewed By: spatel
Differential Revision: https://reviews.llvm.org/D136806
8adfa29706