[Bug c++/66584] gcc differs in static, branch-prediction cost from icc in switch.

["jmcguiness at liquidcapital dot com" <gcc-bugzilla at gcc dot gnu dot org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66584

--- Comment #1 from Jason McG <jmcguiness at liquidcapital dot com> ---
(In reply to Jason McG from comment #0)

I got my static bp summaries wrong, corrected:

> void foo(int i) {
>   switch(i) {
>   case 1:
    bar1(); // gcc: less likely (same as default) | icc: most likely
>     break;
>   case 2:
    bar2(); // gcc: most likely | icc: less likely
>     break;
>   default:
    bar3(); // gcc: less likely (same as case 1) | icc: least likely
>   }
> }