memory_order_acquire:All writes in other threads that release the same atomic variable are visible in the current thread 初步理解为memory_order_consume会比memory_order_acquire少一些强制依赖关系,或者理解为memory_order_acquire的效果是释放操作后有内存栅设置。 gcc的解释更清晰一些: __ATOMIC_RELAXED No ...