This split will ensure we are testing only what a specific optimization is doing
rather than all together.
NOTE: The way I split up the tests is I first split up each of the tests by
subject area by hand that I thought were specifically testing one pass. Then any
tests where the FileCheck tests started to fail due to us not running the other
passes, I put back a copy in the original semantic-arc-opts.sil to ensure we do
not regress.
Note, the tests that I copied for each of these passes
are originally from semantic-arc-opts.sil. I left them there as well so we could
see all of the opts together. I also only copied the ones that were testing pass
specific functionality.
