https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118395
Jonathan Wakely changed:
What|Removed |Added
Assignee|unassigned at gcc dot gnu.org |redi at gcc dot gnu.org
Last
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118395
Jonathan Wakely changed:
What|Removed |Added
Target Milestone|--- |16.0
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118395
--- Comment #4 from Jonathan Wakely ---
Patch posted:
https://gcc.gnu.org/pipermail/gcc-patches/2025-January/673200.html
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118395
--- Comment #3 from Jonathan Wakely ---
So all we need to do is add a single point of contention that all threads
synchronize on every time they call arrive, to make our contention-avoiding
tree barrier work.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118395
--- Comment #2 from Jonathan Wakely ---
Making this constexpr requires major changes:
size_t const __count = (_M_expected + 1) >> 1;
_M_state = std::make_unique<__state_t[]>(__count);
We can't allocate an array during constan
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118395
--- Comment #1 from Jonathan Wakely ---
We also have:
static constexpr ptrdiff_t
max() noexcept
{ return __PTRDIFF_MAX__; }
But this seems wrong, because the constructor does:
size_t const __count = (_M_expected + 1
