]>
git.sesse.net Git - stockfish/log
Marco Costalba [Mon, 5 Jul 2010 12:22:53 +0000 (14:22 +0200)]
Move SplitPoint array under its thread
And cleanup / rename that part of code.
No functional change also with faked split.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 3 Jul 2010 05:02:06 +0000 (06:02 +0100)]
Tweak non-captures scoring
Tested with Orig set at
f5ef5632f so to evaluate
direct gain against 1.8
After 3239 games at 10"+0.1
Mod vs Orig +701 =1906 -632 +7 ELO
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Sun, 4 Jul 2010 19:09:40 +0000 (22:09 +0300)]
Fix TT documentation
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 3 Jul 2010 18:04:05 +0000 (19:04 +0100)]
Reintroduce GCC/ICC rounding hack
Unfortunatly the source of this issue is not in the
different handling of log(0) illegal value.
No functional change on MSVC.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Sat, 3 Jul 2010 16:28:51 +0000 (19:28 +0300)]
Fix Makefile for HPUX
On hpux there is no prefetch.
Reported by Richard Lloyd
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 3 Jul 2010 15:02:07 +0000 (16:02 +0100)]
Fix DIVIDE BY ZERO exception in init_search()
It happens that when d == 0 we calculate:
log(double(0 * 0) / 2)
Unfortunately, log(0) is "illegal" and can generate either a
floating point exception or return a nonsense "huge" value
depending on the platform.
This fixs in the proper way the GCC/ICC rounding difference,
bug was from our side, not in the intel compiler.
Also fixed some few other warnings.
Bug spotted by Richard Lloyd.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 3 Jul 2010 04:48:43 +0000 (05:48 +0100)]
PSQT access functions can be static
Also renamed history access value in something more
in line with the meaning.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 25 Jun 2010 21:55:52 +0000 (22:55 +0100)]
Use only history to score non-captures
It seems there is absolutely no difference in using gains.
After 7025 games at 5"+0
Mod vs Orig +1903 =3236 -1886 (+1 ELO)
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 3 Jul 2010 04:34:57 +0000 (05:34 +0100)]
Restore development version
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 2 Jul 2010 05:53:04 +0000 (06:53 +0100)]
Revert "Another PSQT tuning round"
At longer TC of 1'+0" patch fails:
Orig - Mod: 841 - 819 (-6 elo!)
Just before the release ;-)
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 2 Jul 2010 05:23:15 +0000 (06:23 +0100)]
Stockfish 1.8
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 1 Jul 2010 21:21:09 +0000 (22:21 +0100)]
Mark CheckInfo c'tor as explicit
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 1 Jul 2010 21:18:53 +0000 (22:18 +0100)]
Move singleEvasion assignment out of move's loop
We don't need to recheck after every move.
Spotted by Ralph Stoesser.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Wed, 17 Mar 2010 06:32:26 +0000 (08:32 +0200)]
Another PSQT tuning round
This time with a new algorithm by Joona.
It works basically like this:
repeat
{
1) pick 8000 random positions from qsearch
2) "go depth 8" to get the true evaluation.
3) "eval" to get the stand pat score
4) Adjusting parameters one by one to minimize deltasum between
true evaluation and stand pat scores.
}
* Good news: method seems to converge
* Bad news: Point where it converges is not optimum.
So it's more or less trial and error... sometimes works, sometimes
doesn't. It can give you the right direction, but if you let it run
too long, it fails. Far from scientific ;)
After 14800 games with 5s/game
Orig - Mod: 3318 - 3570 - 7626 (+6 elo)
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 29 Jun 2010 11:05:19 +0000 (13:05 +0200)]
Remove a redundant check in passed pawn eval
When first condition is met then second one is
always true.
Spotted by Ralph Stoesser.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 29 Jun 2010 10:14:44 +0000 (12:14 +0200)]
Code style triviality in san.cpp
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 28 Jun 2010 07:20:34 +0000 (08:20 +0100)]
Rename TranspositionTable 'writes' in 'overwrites'
Better documents what that variable means.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 27 Jun 2010 14:13:22 +0000 (15:13 +0100)]
Extract only exact scores to get the PV
This should allow to skip overwritten nodes because
only in PV we store in TT with VALUE_TYPE_EXACT flag.
Test result for the whole series is:
After 3627 games at 5"
Mod vs Orig +1037 =1605 -985 +5 ELO
After 1311 games at 1'+0"
Mod vs Orig +234 =850 -227 +2 ELO
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 26 Jun 2010 14:22:13 +0000 (15:22 +0100)]
Retire update_pv() and sp_update_pv()
Expand inline instead.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 26 Jun 2010 14:05:38 +0000 (15:05 +0100)]
Retire pv[] from SearchStack
Extract PV info from TT instead of using
a set of arrays. This is almost equivalent
except for cases when TT is full and the PV entry
is overwritten, but this is very rare.
(Almost) No functional change
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 26 Jun 2010 13:42:44 +0000 (14:42 +0100)]
Introduce bestMove to store PV move
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 26 Jun 2010 09:43:59 +0000 (10:43 +0100)]
Cleanup code that stores score in TT
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Sat, 26 Jun 2010 08:12:32 +0000 (11:12 +0300)]
Fix some wrong documentation
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Sat, 26 Jun 2010 08:09:06 +0000 (11:09 +0300)]
Remove unused constant
Fixes warning on ICC
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 25 Jun 2010 17:56:42 +0000 (18:56 +0100)]
Stockfish 1.8 beta 2
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Fri, 25 Jun 2010 17:49:36 +0000 (20:49 +0300)]
Hack to fix GCC/ICC rounding difference
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 24 Jun 2010 16:56:45 +0000 (17:56 +0100)]
Stockfish 1.8 beta 1
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 24 Jun 2010 17:00:03 +0000 (18:00 +0100)]
Revert "Do IID also when we already have a ttMove"
Joona's testing reports very bad results at 5s, 30s and
even 1 minute TC, so revert.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 5 Jun 2010 17:05:29 +0000 (18:05 +0100)]
Call apply_weight() for both colors in one go
Due to rounding errors in apply_weight() where we divide
by 0x100 it is not possible to keep some functionality.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 13 Jun 2010 10:05:04 +0000 (11:05 +0100)]
Tweak unstoppable pawns detection
A pawn is unstoppable also if enemy king can reach it
but path to queening is protected.
Original idea by Ralph Stoesser fixed by me.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 21 Jun 2010 19:50:20 +0000 (20:50 +0100)]
Merge branch 'master' of free2.projectlocker.com:sf
Tord Romstad [Mon, 21 Jun 2010 09:25:06 +0000 (11:25 +0200)]
Moved a misplaced #endif in misc.cpp, which broke compilation in
Mac OS X.
No functional change.
Marco Costalba [Sat, 19 Jun 2010 15:48:36 +0000 (16:48 +0100)]
Test killer for legality earlier
Many killers moves, around 40%, are not legal, so
skip earlier in this case.
Some Movepicker c'tor cleanup while there.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 19 Jun 2010 14:44:03 +0000 (15:44 +0100)]
Optimize for king moves in see_sign()
Because we only test legal moves, a king move
cannot have negative SEE.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 19 Jun 2010 10:10:54 +0000 (11:10 +0100)]
Move prefetch() out of TT
This code is platform specific and has nothing to
do with TT class, so move to misc.cpp
This patch is a prerequisite to use extend prefetch use
also to other hash tables apart from Transposition Table.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 12 Jun 2010 13:47:30 +0000 (14:47 +0100)]
Extend checks only if SEE is non-negative
Idea from Dr. Hyatt
After 10k games at 5"+0 on my QUAD
Mod vs Orig +2750 =4601 -2649 +4 ELO
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 13 Jun 2010 01:26:43 +0000 (02:26 +0100)]
Revert "Use ply counter in Position object"
Search ply and game ply are rwo different things !
Revert bogus commit.
No functional change on bench, but it changes in real games
when engine sends all the moves up to current one.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 12 Jun 2010 09:29:50 +0000 (10:29 +0100)]
Set LSNTime to 100 ms
This is a timeout compatible with very short TC of 5 sec/game.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 5 Jun 2010 17:04:58 +0000 (18:04 +0100)]
Remove a wrong FIXME
If we are there it means we already had that info stored in TT,
so we don't need to overwrite with the same content !
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 5 Jun 2010 09:50:24 +0000 (11:50 +0200)]
Avoid a double copy when saving a TTEntry
In statement:
*tte = TTEntry(posKey32, v, t, d, m, generation, statV, kingD);
We first create a TTEntry, then we copy the temporary entry to
its final destination in *tte then we discard the TTEntry.
Instead of this assign the fields directly to the destination TTEntry.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 5 Jun 2010 07:40:47 +0000 (09:40 +0200)]
Avoid calling evaluate() while razoring
Micro optimization that gives a 0.5% speed improvment
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 4 Jun 2010 07:46:38 +0000 (09:46 +0200)]
Offset pv[] always from 0
We don't need to offset from current ply.
Also rewritten a bit update_pv()
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 3 Jun 2010 17:29:48 +0000 (18:29 +0100)]
Big qsearch() cleanup
It is more clear and should be a bit faster too.
Reverted also previous optimization patch because
seems do not increase actual speed.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 2 Jun 2010 10:47:53 +0000 (11:47 +0100)]
Extend intermediate LMR to root search
Almost no change, but it is in sync with what we do in search
and in any case the ELO difference is very small (because the
events when the intermediate research triggers are very rare),
too small to be measured, we just verify we don't have any
unexpected regressions.
After 802 games at 1+0 full QUAD
Mod vs Orig +114 =581 -107 +3 ELO
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 3 Jun 2010 10:10:12 +0000 (12:10 +0200)]
Shortcut futility pruning in qsearch
If we have pruned one capture due to its final value
we can prune also following ones because captures are
MVV ordered.
Also avoid a compare when not in PV because in that
case is always false.
No functional change.
Marco Costalba [Wed, 2 Jun 2010 19:37:18 +0000 (20:37 +0100)]
Do not pass threadID as argument of search() and evaluate()
Get it from the position instead.
A good semplification of function calling and a speedup too.
No functional change also with faked split.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 2 Jun 2010 19:13:51 +0000 (20:13 +0100)]
Save threadID info in Position
This is the best place because when we split we do a
copy of the position and also threadID, once set in a
given position, never changes anymore.
Forbid use of Position's default and copy c'tor to avoid
nasty bugs in case a position is created without explicitly
setting the threadID.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 2 Jun 2010 13:18:23 +0000 (14:18 +0100)]
Don't initialize excludedMove and skipNullMove at each node
Do it once at the beginning becuase they are always reset
after use in the calling place where are set.
No functional change also with faked split.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 2 Jun 2010 12:32:31 +0000 (13:32 +0100)]
Use SearchStack to pass allowNullmove
Also renamed allowNullmove in skipNullMove to reverse
the logic so that the field is initialized to 0 (false)
instead of 1 (true).
No functional change also with faked split.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 2 Jun 2010 12:22:48 +0000 (13:22 +0100)]
Use SearchStack to pass excludedMove
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 2 Jun 2010 10:25:02 +0000 (11:25 +0100)]
Fix white space breakage
No functional change with faked split.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Wed, 2 Jun 2010 09:49:34 +0000 (10:49 +0100)]
Sync sp_search() with main search()
And fix qserahc() dispatch also there.
No functional change tested wit Faked Split.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 1 Jun 2010 20:28:45 +0000 (21:28 +0100)]
Retire bitScanReverse32()
Use log() instead because we are not in speed
critical paths.
Also a bit of renaming and code shuffle while there.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 1 Jun 2010 18:49:46 +0000 (19:49 +0100)]
Retire ok_to_do_nullmove()
Has been remained the same from ages also with the FIXME.
Retire for now and rearrange the conditions order for
maximum performance.
Also a small touch at null zugzwang verification.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 1 Jun 2010 14:31:24 +0000 (15:31 +0100)]
Move invariant of singular ext. check out of loop
It is almost always false.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 1 Jun 2010 14:02:14 +0000 (15:02 +0100)]
Rename ok_to_prune() in connected_threat()
It is more up to the point. Also small speedup
due to checking for threat move before calling
the function. This saves more then 90% of function calls.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 1 Jun 2010 12:03:46 +0000 (13:03 +0100)]
Avoid double function dispatch
In 44% of cases we call search() just to call
qsearch() one moment later, avoid this double dispatch.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 1 Jun 2010 07:21:11 +0000 (08:21 +0100)]
Retire init_node()
Also don't poll in qsearch()
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Tue, 1 Jun 2010 06:36:30 +0000 (07:36 +0100)]
Don't init SplitPointStack[i][j].parent
It is already set to zero because is allocated in
the global storage area.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 31 May 2010 20:38:34 +0000 (21:38 +0100)]
Use ply counter in Position object
And avoid a redundant one passed as argument in
search calls.
Also renamed gamePly in ply to better clarify this
is used as search ply and is set to zero at the
beginning of the search.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 31 May 2010 17:32:00 +0000 (18:32 +0100)]
Fix SearchStack and ply misalignment in RootMoveList
In RootMoveList c'tor we call qsearch() with
ply == 1 but SearchStack at 0.
We never noticed before because in qsearch we don't access
previous's ply SearchStack, otherwise we would have got
a nice crash ;-)
This bug is a fall down of previous patch.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 30 May 2010 07:56:38 +0000 (08:56 +0100)]
Convert SearchStack ss[] to SearchStack*
Use a pointer to current SearchStack to avoid ss[ply]
address calculation.
Gives 1% speedup on Intel compiler
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 31 May 2010 10:05:04 +0000 (12:05 +0200)]
Add function to get ply from position
It will be used by future patches.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 31 May 2010 09:47:15 +0000 (11:47 +0200)]
Move gamePly among the StateInfo data
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 31 May 2010 11:01:33 +0000 (12:01 +0100)]
Move promotion and ep under pawn handling
And remove from main do_move() flow. Just a small speedup
because we avoid two branches in the common case.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 31 May 2010 05:13:34 +0000 (06:13 +0100)]
Thread::splitPoint is a volatile pointer
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 29 May 2010 09:01:25 +0000 (10:01 +0100)]
Do IID also when we already have a ttMove
In case tte->depth() is far lower the current depth
and we are in a PV node.
Almost 45% of researches give a different ttMove !
After 999 games at 1+0
Mod vs Orig +174 =694 -131 +15 ELO !!!!!!!
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 28 May 2010 22:31:05 +0000 (23:31 +0100)]
If LMR search fails high research at intemediate depth
Do not search immediately at full depth, but try a second
chance at lower depth. This is a feature that should scale
well because become important at high depths where we have
big reductions and also big savings in avoiding a costly
full depth search.
After 942 games at 1+0
Mod vs Orig +158 =645 -139 +7 ELO
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 29 May 2010 08:43:18 +0000 (09:43 +0100)]
Fix IIDMargin description
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 29 May 2010 07:54:15 +0000 (08:54 +0100)]
Retire zobMaterial[]
Use zobrist[] array to compute material key.
Space save of 2KB in L1 cache.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 28 May 2010 11:21:03 +0000 (12:21 +0100)]
Change zobMaterial[] index 0 definition
The index at 0 was reserved for no-pieces
information. But we don't need that.
This is a prerequisite for next patch.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 29 May 2010 07:11:36 +0000 (08:11 +0100)]
Use Key type instead of Bitboard
They are both 64 bits unsigned integer, but it
is correct to use the proper type.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 23 May 2010 13:51:15 +0000 (14:51 +0100)]
Revert non-linear threats evaluation
After 999 games at 1+0
Mod vs Orig +148 =712 -139 +3 ELO
The added complexity doesn't seems to pay off and could
even scale worst with longer TC. So revert.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 22 May 2010 18:14:19 +0000 (19:14 +0100)]
Remove an useless assert in evaluate_passed_pawns()
The tested square comes from a bitboard anded with
pos.pieces_of_color(Us), so assert is useless.
Another nitpick report by Marek Kwiatkowski ;-)
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 22 May 2010 17:38:45 +0000 (18:38 +0100)]
Account for double pawn push in evaluate_unstoppable_pawns()
One of the most nitpicking patches I have ever seen.
Of course almost no functional change, but added just
becasue we are very pedantic ;-)
Spotted by Marek Kwiatkowski
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 17 May 2010 05:08:47 +0000 (06:08 +0100)]
Fix StormOpenFileBonus[] bug
It was erroneusly reversed. Bug from Glaurung times.
Probably a full re-tuning is needed anyhow.
Spotted and fixed by Ralph Stoesser.
After 999 games almost no change, but modified anyway
for documentation reasons.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 22 May 2010 11:45:05 +0000 (12:45 +0100)]
Fix a warning on array of size 0 under Windows
And better document new reality.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Sat, 22 May 2010 08:34:35 +0000 (11:34 +0300)]
Drop completely illogical ei.kingDanger == 0 requirement
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Sat, 22 May 2010 08:56:46 +0000 (11:56 +0300)]
Always save static value and kingDanger to TT
Around 5% speed-up
No functional change
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Sat, 22 May 2010 08:51:22 +0000 (11:51 +0300)]
Drop TTClusterSize from 5 to 4
Very small obvious functional change
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 22 May 2010 10:55:01 +0000 (11:55 +0100)]
Add non-linear threats evaluation
Increase threats score according to the number of
threats and to the side to move.
Constants have been balanced after ~34k iterations.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 22 May 2010 10:35:15 +0000 (11:35 +0100)]
Revert "Simple implementation of strong YBWC"
It does not seem to increase anything even with a QUAD,
so revert.
After 1000 games with a QUAD
Mod - Orig: 500 - 497 (+1 elo)
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Fri, 21 May 2010 19:17:59 +0000 (22:17 +0300)]
Add some automatic detection for Windows
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Fri, 21 May 2010 05:31:27 +0000 (08:31 +0300)]
Documentation fix
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Fri, 21 May 2010 05:05:36 +0000 (08:05 +0300)]
Correct 'prefetch' handling for Makefile
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 20 May 2010 20:37:37 +0000 (21:37 +0100)]
Let prefetch to be enabled by default on Windows
When compiling with MSVC we don't use the Makefile
so tweak a bit the Makefile to allow to let prefetch
in by default so that it works under Windows.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Thu, 20 May 2010 15:24:45 +0000 (18:24 +0300)]
Small tweaks to install gcc-profile-clean targets
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Wed, 19 May 2010 19:44:26 +0000 (22:44 +0300)]
Modify source to follow new Makefile
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Joona Kiiski [Wed, 19 May 2010 17:37:56 +0000 (20:37 +0300)]
Rewrite Makefile
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 13 May 2010 08:06:27 +0000 (10:06 +0200)]
Simple implementation of strong YBWC
No gain in the worst case of 2 threads, but also no
loss and good potential for QUAD or OCTAL machines.
After 922 games at 1+0 with 2 threads
Mod vs Orig +143 =533 -143 +0 ELO
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Thu, 20 May 2010 16:48:19 +0000 (17:48 +0100)]
Remove an assert in evaluate_passed_pawns()
We already tested few lines before with:
assert(pos.pawn_is_passed(Us, s));
Spotted by Marek Kwiatkowski.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Mon, 17 May 2010 06:04:24 +0000 (07:04 +0100)]
Indentation fix in middle-game evaluation
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 15 May 2010 12:29:21 +0000 (13:29 +0100)]
Introduce init_attack_tables() in evaluate()
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 15 May 2010 12:00:32 +0000 (13:00 +0100)]
Do not consider discovered checks in king safety
Does not help and it slows downs a bit because it
is not cheap to get the possible discovered checks
out of a position.
After 997 games at 1+0
Mod vs orig +153 =692 -152 +0 ELO
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sun, 16 May 2010 07:56:58 +0000 (08:56 +0100)]
Fix RootMove::operator<() description
Reported by Melvin Sprague.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 15 May 2010 23:56:05 +0000 (00:56 +0100)]
Remove an useless Max() in passed pawns evaluation
There is no reason for that since tr cannot become negative.
Spotted by Ralph Stoesser.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Sat, 15 May 2010 20:44:57 +0000 (21:44 +0100)]
Threath tuning results
Final values for threath tuning (after ~30k iterations)
Verified to be equivalent with tuning branch.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 14 May 2010 16:12:10 +0000 (17:12 +0100)]
Add support for gcc-profile
It's now possible to build PGO builds with GCC
Patch from Oystein Johansen
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 14 May 2010 13:11:32 +0000 (15:11 +0200)]
Rename OutpostMask[] in AttackSpanMask[]
This is a more standard naming (see chessprogramming wiki)
and is more stick to what table is and not what is used for.
Code in pawns.cpp is a bit more readable now, at least for me ;-)
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>
Marco Costalba [Fri, 14 May 2010 11:51:29 +0000 (12:51 +0100)]
Micro optimization in evaluate_pawns()
Avoid a double bitcount in test for candidate passed
pawn when we don't have any supporting pawn.
Also use outpost_mask() instead of build it up on
the fly.
No functional change.
Signed-off-by: Marco Costalba <mcostalba@gmail.com>