>Hi Simon,
>
>He is going to do some investigation in the methods and
>write down the possibilities and then he is going to implement
>something from that for PostgreSQL.
>
>When will this work be complete? We are days away from completing main
>work on 8.4, so you won't get much discussion on this for a few months
>yet. Will it be complete in time for 8.5? Or much earlier even?
>
>The first guess is that the work will be done for 8.6.
>Dano is supposed to finish the work and defend his thesis in something
a bit more than a year.
>
>Julius, you don't mention what your role is in this. In what sense is
>Dano's master's thesis a "we" thing?
>
>I am Dano's mentor and we have a closed contact with Zdenek as well.
>We would like the project to become a "we" thing as another reason why
to work on the project.
>It seems to be better to research some ideas at the begging and discuss
the stuff
>during development than just individually writing some piece of code
which could
>be published afterwards. Especially, when this area seems to be of
interest of
>more people.
Threads are where future performance is going to come from:
General purpose->
http://www.setup32.com/hardware/cpuchipset/32core-processors-intel-reach
e.php
GPU->
http://wwwx.cs.unc.edu/~lastra/Research/GPU_performance.html
http://www.cs.unc.edu/~geom/GPUSORT/results.html
Database engines that want to exploit the ultimate in performance will
utilize multiple threads of execution.
True, the same thing can be realized by multiple processes, but a
process is more expensive than a thread.