[Bug middle-end/64928] [5/6/7/8 Regression] Inordinate cpu time and memory usage in "phase opt and generate" with -ftest-coverage -fprofile-arcs

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64928

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|2015-02-09 00:00:00         |2017-8-21

--- Comment #24 from Richard Biener <rguenth at gcc dot gnu.org> ---
I think we made LIM more powerful and/or changed coalescing.  We didn't really
address the issue with LIMs lack of a costmodel or out-of-SSA coalescing being
quadratic in size requirement (but I see RTL passes bumping up to 5GB of memory
use as well, REE for example, also noted elsewhere).