Re: Patch: fix lock contention for HASHHDR.mutex
| От | Amit Kapila |
|---|---|
| Тема | Re: Patch: fix lock contention for HASHHDR.mutex |
| Дата | |
| Msg-id | CAA4eK1KOnADjGV2D0EiW2jUowp7uQp3MQmrR4wQY+P6it5m8ow@mail.gmail.com обсуждение исходный текст |
| Ответ на | Re: Patch: fix lock contention for HASHHDR.mutex (Aleksander Alekseev <a.alekseev@postgrespro.ru>) |
| Ответы |
Re: Patch: fix lock contention for HASHHDR.mutex
Re: Patch: fix lock contention for HASHHDR.mutex |
| Список | pgsql-hackers |
On Thu, Dec 17, 2015 at 9:33 PM, Aleksander Alekseev <a.alekseev@postgrespro.ru> wrote:
>
> > It'd really like to see it being replaced by a queuing lock
> > (i.e. lwlock) before we go there. And then maybe partition the
> > freelist, and make nentries an atomic.
>
> I believe I just implemented something like this (see attachment). The
> idea is to partition PROCLOCK hash table manually into NUM_LOCK_
> PARTITIONS smaller and non-partitioned hash tables. Since these tables
> are non-partitioned spinlock is not used and there is no lock
> contention.
>
>
> > It'd really like to see it being replaced by a queuing lock
> > (i.e. lwlock) before we go there. And then maybe partition the
> > freelist, and make nentries an atomic.
>
> I believe I just implemented something like this (see attachment). The
> idea is to partition PROCLOCK hash table manually into NUM_LOCK_
> PARTITIONS smaller and non-partitioned hash tables. Since these tables
> are non-partitioned spinlock is not used and there is no lock
> contention.
>
This idea can improve the situation with ProcLock hash table, but I
think IIUC what Andres is suggesting would reduce the contention
around dynahash freelist and can be helpful in many more situations
including BufMapping locks.
В списке pgsql-hackers по дате отправления: