Re: Parallel Index Scans

Haribabu Kommi <kommi.haribabu@gmail.com>

From: Haribabu Kommi <kommi.haribabu@gmail.com>
To: Amit Kapila <amit.kapila16@gmail.com>
Cc: Robert Haas <robertmhaas@gmail.com>, Anastasia Lubennikova <a.lubennikova@postgrespro.ru>, Anastasia Lubennikova <lubennikovaav@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>, Rahila Syed <rahilasyed.90@gmail.com>
Date: 2017-01-27T01:23:32Z
Lists: pgsql-hackers
On Mon, Jan 23, 2017 at 5:07 PM, Amit Kapila <amit.kapila16@gmail.com>
wrote:

> On Fri, Jan 20, 2017 at 7:29 AM, Haribabu Kommi
> <kommi.haribabu@gmail.com> wrote:
> >
> > On Thu, Jan 19, 2017 at 1:18 AM, Amit Kapila <amit.kapila16@gmail.com>
> > wrote:
> >> >
> >> > +extern BlockNumber _bt_parallel_seize(IndexScanDesc scan, bool
> >> > *status);
> >> > +extern void _bt_parallel_release(IndexScanDesc scan, BlockNumber
> >> > scan_page);
> >> >
> >> > Any better names for the above functions, as these function will
> >> > provide/set
> >> > the next page that needs to be read.
> >> >
> >>
> >> These functions also set the state of scan.  IIRC, these names were
> >> being agreed between Robert and Rahila as well (suggested offlist by
> >> Robert).  I am open to change if you or others have any better
> >> suggestions.
> >
> >
> > I didn't find any better names other than the following,
> >
> > _bt_get_next_parallel_page
> > _bt_set_next_parallel_page
> >
>
> I am not sure using *_next_* here will convey the message because for
> backward scans we set the last page.  I am open to changing the names
> of functions if committer and or others prefer the names suggested by
> you.


OK. I am fine with it.
I don't have any other comments on the patch.

Regards,
Hari Babu
Fujitsu Australia

Commits

  1. Replace min_parallel_relation_size with two new GUCs.

  2. Factor out logic for computing number of parallel workers.

  3. Support condition variables.

  4. Allow parallel custom and foreign scans.

  5. Add a C API for parallel heap scans.