Re: [gnugo-devel] arend_1_26.5: owl callback fix

[Daniel Bump]

> > for all_batches was 4.45 % and the speedup for owl.tst
> > was 2.29 %. There could be some randomness in the timing
> > results but I'm convinced this patch is a good speedup.
> Hmm. I measured 16% for nngs.tst with gprof. I think gprof's accuracy
> should be quite good; also, this agrees with the save of reading nodes.
> However, gprof doesn't measure time spent in system calls. Maybe my
> handling of the pattern list memory is inefficient. I will do some more
> tests about this.

When I ran those tests I had two GNU Go processes running at
the same time. Maybe that invalidated the timing results. I
hoped that on a SMP machine it would be OK.

Last night I ran the regressions with the current CVS and only
one process this time with the options:

configure option enabled: experimental influence
configure option enabled: experimental connections
configure option enabled: alternate connections

The results are the same as yesterday except we lost the
unexpected PASS for arend:28 for some reason.

The good news is that the tests finished in 28108 seconds = 8 hours,
a pretty good time. I'll run them again with the experimental
influence turned off. I think we're all eager to make the experimental
influence standard. Maybe we've gained enough speed that we can
consider doing that.

Dan