https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114589
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Blocks| |85316
--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
So when doing the sinking it's too late since there's no VRP between early
sinking and late unrolling and also threadfull doesn't figure out
<bb 3> [local count: 105119324]:
_11 = &o_3(D)->val;
_12 = 1;
_13 = 4;
_10 = _11 + 4;
ivtmp.11_14 = (unsigned long) _11;
<bb 4> [local count: 955630224]:
# ivtmp.11_17 = PHI <ivtmp.11_2(6), ivtmp.11_14(3)>
_18 = (void *) ivtmp.11_17;
i_6 = MEM[(const int *)_18];
f (i_6);
ivtmp.11_2 = ivtmp.11_17 + 4;
__for_begin_19 = (const int *) ivtmp.11_2;
if (__for_begin_19 != _10)
goto <bb 6>; [89.00%]
else
goto <bb 5>; [11.00%]
<bb 6> [local count: 850510900]:
goto <bb 4>; [100.00%]
but IIRC it doesn't consider the loop entry "fallthru" as "threading" source.
Nevertheless I have a fix for the sinking issue.
Referenced Bugs:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85316
[Bug 85316] [meta-bug] VRP range propagation missed cases
