Hi,
On Fri, Oct 25 2024, Richard Biener wrote:
> The following places a few strathegic asserts so we do not end up
> with COND_EXPRs with a comparison as the first operand during
> vectorization.
>
> Bootstrap and regtest running on x86_64-unknown-linux-gnu.
>
> These are the asserts I'd like to keep, I will followup with quite
> some dead code removal when this holds.
>
> * tree-vect-slp.cc (vect_get_operand_map): Mark
> COMPARISON_CLASS_P COND_EXPR condition path unreachable.
> * tree-vect-stmts.cc (vect_is_simple_use): Likewise.
> (vectorizable_condition): Assert the COND_EXPR condition isn't
> COMPARISON_CLASS_P.
> ---
> gcc/tree-vect-slp.cc | 2 +-
> gcc/tree-vect-stmts.cc | 8 ++------
> 2 files changed, 3 insertions(+), 7 deletions(-)
>
> diff --git a/gcc/tree-vect-slp.cc b/gcc/tree-vect-slp.cc
> index b192328e3eb..2e98a943e06 100644
> --- a/gcc/tree-vect-slp.cc
> +++ b/gcc/tree-vect-slp.cc
> @@ -549,7 +549,7 @@ vect_get_operand_map (const gimple *stmt, bool
> gather_scatter_p = false,
> {
> if (gimple_assign_rhs_code (assign) == COND_EXPR
> && COMPARISON_CLASS_P (gimple_assign_rhs1 (assign)))
> - return cond_expr_maps[swap];
> + gcc_unreachable ();
This seems to have been the only use of cond_expr_maps which we still
have, however. Should the array be removed too?
Thanks,
Martin
