https://gcc.gnu.org/bugzilla/show_bug.cgi?id=121712
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Last reconfirmed| |2025-08-29
Ever confirmed|0 |1
Status|UNCONFIRMED |NEW
CC| |rguenth at gcc dot gnu.org
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Estimating # of iterations of loop 1
Analyzing # of iterations of loop 1
exit condition [a_9(D) + 4, + , 4](no_overflow) != b_8(D)
bounds on difference of bases: -18446744073709551615 ... 18446744073709551611
result:
# of iterations (((unsigned long) b_8(D) - (unsigned long) a_9(D)) - 4)
/[ex] 4, bounded by 4611686018427387902
Statement (exit)if (b_8(D) != a_12)
is executed at most (((unsigned long) b_8(D) - (unsigned long) a_9(D)) - 4)
/[ex] 4 (bounded by 4611686018427387902) + 1 times in loop 1.
Analyzing # of iterations of loop 1
exit condition [a_9(D) + 4, + , 4](no_overflow) != b_8(D)
bounds on difference of bases: -18446744073709551615 ... 18446744073709551611
result:
# of iterations (((unsigned long) b_8(D) - (unsigned long) a_9(D)) - 4)
/[ex] 4, bounded by 4611686018427387902
Loop 1 iterates at most -2 times.
Loop 1 likely iterates at most -2 times.
wut? (probably a dump issue)
niter analysis uses it's simplify_using_initial_conditions but that hasn't
seen any TLC "recently". The initial conditions include
<bb 2> [local count: 347387056]:
_1 = b_8(D) - a_9(D);
if (_1 == 4)
goto <bb 5>; [34.00%]
We could possibly do a walk_tree on the niter expression, trying to compute
a singleton range for sub-expressions, or even just try on the whole
expression.
