Re: [HACKERS] Postgres Speed or lack thereof

>
> Can someone remind me where we left this memory context stuff?
>

    I've  changed  AllocSetAlloc()  and  friends  to manage small
    allocations in bigger blocks together  where  the  sizes  are
    round up to powers of two and free's chunks in the blocks are
    reused in place. The entire blocks get only free'd when  they
    get completely free.

    Then  I've  changed  palloc() etc. into macros (which started
    the discussion on mcxt.h - sorry).

    Overall  speedup  from  the  changes  (for   regression)   is
    something about 10-15%.


Jan

--

#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me.                                  #
#======================================== jwieck@debis.com (Jan Wieck) #