On Mon, Apr 4, 2011 at 7:04 PM, Robert Haas <robertmh...@gmail.com> wrote:
> On Mon, Apr 4, 2011 at 7:16 AM, Alexander Korotkov <aekorot...@gmail.com> > wrote: > > Project name > > Fast GiST index build > > Would/could/should this be implemented in a manner similar to the > existing "GIN fast update" feature? > I've mentioned this problem in item #2 in project details. In short. Problem is in concurrent selects. Buffers size is significant and their scan in concurrent select can cause significant slow down. Probably, compromise can be achived by using for smaller buffers or something like this, but it's topic of separate research. It doesn't seems to be feasible for me to give a production solution of this problem during GSoC. ---- With best regards, Alexander Korotkov.
