On Mon, Jun 05, 2006 at 11:02:33PM -0400, Tom Lane wrote:
> Just got this rather surprising result:
<snip bogus explain output>
> The "Total runtime" is correct AFAICT, which puts the top node's "actual
> time" rather far out in left field. This is pretty repeatable on my old
> slow HPPA machine. A new Xeon shows less of a discrepancy, but it's
> still claiming top node actual > total, which is not right.
Wierd. Can you get the output of *instr in each call of
InstrEndLoop? Preferably after it does the calculation but before it
clears the values. So we get an idea of number of samples and what it
guesses. SampleOverhead would be good too.
I know my version produced sensible results on my machine and the
handful of people testing, so I'll try it again with your changes, see
how it looks...
Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
> From each according to his ability. To each according to his ability to litigate.