Thomas Lockhart <lockhart@alumni.caltech.edu> writes:
> Another followup: on 7.0.2, with different optimizations etc,
> sum(float8) takes 1.95 seconds, rather than the 5.2 on the current tree.
> I'd better look at the compilation optimizations; is there another
> explanation for the factor of 2.6 difference (!!)?
If you are running with --enable-cassert then there is a whole bunch
of memory-stomp debugging overhead turned on in current sources,
including such time-consuming stuff as clearing every pfree'd block.
7.0.*'s --enable-cassert is not nearly as expensive.
I plan to make that stuff not-default when we go beta, but right now
it seems like a good idea to have it on for testing...
regards, tom lane
