https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114030
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Component|tree-optimization |middle-end
--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
FRE only tries to optimize same sized "punned" loads, it doesn't try to handle
truncation. At -O2 it looks like some missed SRA (some inlining only happens
at IPA time with -O2).
In the end it's maybe better suited for RTL optimization.
