On Sat, Mar 25, 2006 at 04:24:05PM +0100, Martijn van Oosterhout wrote:
> I agree. However, if it's the overhead of calling gettimeofday() that
> slows everything down, perhaps we should tackle that end. For example,
> have a sampling mode that only times say 5% of the executed nodes.
>
> EXPLAIN ANALYZE SAMPLE blah;
<snip>
> You could do a non-random sampling fairly easily:
Actually, I thought of a better way to control the sampling that's
probably better than the crude if-then-else structure I gave which also
takes advantage of the fact that the numbers are floating point:
InstrInit:
next_sample = 0;
InstrStart: if( ntuples < 16 ) dosample = yes; else if( ntuples > next_sample ) { dosample = yes;
next_sample += log2(ntuples); } else dosample = no;
This will sample approxiamtly log2 of the actual executions. You could
use log10(), sqrt() or any other function you find reflects the sample
you want.
Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
> Patent. n. Genius is 5% inspiration and 95% perspiration. A patent is a
> tool for doing 5% of the work and then sitting around waiting for someone
> else to do the other 95% so you can sue them.