https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105706
Kewen Lin <linkw at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |linkw at gcc dot gnu.org
Last reconfirmed| |2022-05-24
Ever confirmed|0 |1
Status|UNCONFIRMED |NEW
--- Comment #1 from Kewen Lin <linkw at gcc dot gnu.org> ---
(In reply to seurer from comment #0)
> g:68e0063397ba820e71adc220b2da0581dce29ffa, r13-707-g68e0063397ba82
> make -k check-gcc RUNTESTFLAGS="powerpc.exp=gcc.target/powerpc/pr78604.c"
> FAIL: gcc.target/powerpc/pr78604.c scan-tree-dump-times vect
> "vect_model_simple_cost" 8
> # of expected passes 4
> # of unexpected failures 1
>
>
> Does this test need to be adjusted, too?
>
Yes. The related commit mainly adjust the below gimple:
iftmp.7_4 = _1 < _2 ? val2_7(D) : val1_8(D);
into
_3 = _1 >= _2;
iftmp.7_4 = _3 ? val1_8(D) : val2_7(D);
So for each function it has two statements (comparison and condition) calling
vect_model_simple_cost instead of one (condition). By checking the history, the
case was updated with r8-731, which meant to verify if vectorized COND_EXPRs
calling vect_model_simple_cost.
One simple fix is to adjust the match count to 16.
