https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119919
Jan Hubicka <hubicka at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Depends on| |119902
--- Comment #3 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
C equivalent is:
int a[9*9];
bool b[9];
void test()
{
for (int i = 0; i < 9; i++)
{
b[i] = a[i*9] != 0;
}
}
there are two problems. First is costing of the open-coded gather. This is
PR119902.
Second problem is that we over-estimate cost of
_7 = sudoku1[_6];
_8 = _7 != 0;
for some reason vectorizer does:
_7 != 0 4 times vector_stmt costs 16 in body
<unknown> 1 times vector_load costs 12 in prologue
_8 ? 1 : 0 4 times vector_stmt costs 64 in body
_8 ? 1 : 0 is redundant (since _8 is already either 1 or 0) and, moreover,
over-costed.
Referenced Bugs:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119902
[Bug 119902] open-coded scatter/gather should not account vec_to_scalar cost
