[Bug middle-end/63155] [5/6/7 Regression] memory hog

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63155

--- Comment #14 from Richard Biener <rguenth at gcc dot gnu.org> ---
The plan is to re-write coalescings conflict graph by using SSA-based liveness
queries according to https://hal.inria.fr/inria-00192219

Well, at least I'd like to experiment with that as it might be useful in other
places as well (cost considerations wrt register pressure changes).

Only experiments will show if we trade too much compile-time for the reduction
in memory use (not computing liveness / conflict graph).