> LGTM, but please note there are a whole lot of CMOVE issues on x86,
> collected in [1] meta-bug, especially [2].
>
> [1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85559
> [2] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56309
Thanks for noting this, I've verified my patch have no impact for t
On Tue, Jan 7, 2025 at 8:37 AM Hongyu Wang wrote:
>
> Hi,
>
> For later processors, the pipeline went deeper so the penalty for
> untaken branch can be larger than before. Add a new parameter
> br_mispredict_scale to describe the penalty, and adopt to
> noce_max_ifcvt_seq_cost hook to allow longer
