[Bug ipa/120120] [16 Regression] gcc-16: performance regression with -O3 compared to gcc-15 since r16-170-ga670ebde399548

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120120

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|target                      |ipa
                 CC|                            |hubicka at gcc dot gnu.org,
                   |                            |jamborm at gcc dot gnu.org

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 15 does

t.c:257:10: optimized:  Inlined codec47_block.constprop/17 into
codec47_comp2/16 which now has time 40289.130615 and size 157, net change of
-15.

while GCC 16 no longer does this.  Adding -fno-ipa-cp to GCC 15 causes
us to recursively inline instead.

But the important difference is likely that with the inlining we get
a lot of complete unrolling but not without.

I can imagine that IPA CP is maybe confused by the frontend no longer
promoting arguments to codec47_block which receives uint16_t size.