On 11/15/25 5:02 AM, Richard Biener wrote:
In principle the dependence analysis part of LIM is quadratic, but I have not
seen actual cases in a long time where it would outcompete FRE or PRE in terms
of compile time (but IIRC we do not put hard limits on the max work it does).
I will look at the actual test fallout then.
But it's only working on loops and likely only on a subset of the
SSA_NAMEs in the loops to prove properties. So even if it's quadratic,
it's probably a lot cheaper than PRE.
jeff
