🚀 go-pugleaf

I have noted some things in search.c and genmoves.c that seem
to be non-optimal compared to earlier versions of UChess:

1) In genmoves.c the sidebit is used in some references to the
killtable (killt[xxx]), It would seem that the
sidebit is not necessary, as Kong Sian pointed out
a while ago.  In any case it should either be used
or NOT used consistently, this is not the case in
genmoves.c.

I think it should not be used ever, any comments?


2) in search() when calculating do_pvs after if (pnt == pbst)
   it is probably correct, and much more efficient to use:

   do_pvs = flag.pvs && PVari; rather than
   do _pvs = flag.pvs && !null && (Prvar[ply == .....

3) I do not believe the pre4pl67 eval stuff is better than
   the newer eval routines, I am sticking with the newer evals.
   Can anyone tell me different?

Overall I do see a performance increase going to 4pl70 ttable
stuff (search/ttable.c etc) vs the old ways, so I think
it is a step forward.


-Roger
--------------------------------------------------------------
bix: ruzun
NET: uzun@crash.cts.com