On Monday, October 29, 2012 09:33:25 PM Tom Lane wrote:
> Christian Kruse <cjk+postgres@defunct.ch> writes:
> > I created a patch which implements MAP_HUGETLB for sysv shared memory
> > segments (PGSharedMemoryCreate). It is based on tests of Tom Lane and
> > Andres Freund, I added error handling, huge page size detection and a
> > GUC variable.
>
> My recollection is we'd decided not to pursue that because of fears that
> it could make performance horribly worse on some platforms.
Hm. I don't so. Maybe youre remembering our discussion about mlock?
MAP_HUGETLB is linux specific anyway, there are other OSes doing this, but I
don't think a standard exists.
http://archives.postgresql.org/message-
id/201206292152.40311.andres%402ndquadrant.com and
http://archives.postgresql.org/message-
id/CA%2BTgmoZGX0Pi0rw5sDH0Uz%3D03WkQ%3DmnoAW3TXVEfvUpyW%2BfMjw%40mail.gmail.com
seem to be the most recent discussions arround this.
The parts I like most about having usable hugepages arround is usable 'ps'
output and way smaller pagetables...
Greetings,
Andres
--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
