Stockfish Testing Queue

Finished - 27208 tests

11-10-15 lb time_loss_fix diff
ELO: -2.43 +-4.3 (95%) LOS: 13.6%
Total: 10000 W: 1991 L: 2061 D: 5948
10000 @ 2+0.01 th 7 I hope to see lots of time losses with lazy_smp here :-)
10-10-15 bi pawn_imbalance diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 8998 W: 1629 L: 1716 D: 5653
sprt @ 15+0.05 th 1 try 3 (fixed)
10-10-15 sn check diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 3811 W: 669 L: 779 D: 2363
sprt @ 15+0.05 th 1 Suppress popcounts in check analysis
10-10-15 bi pawn_imbalance diff
LLR: -2.16 (-2.94,2.94) [0.00,5.00]
Total: 31200 W: 5854 L: 5810 D: 19536
sprt @ 15+0.05 th 1 Try 3, against Knight/King should be winning
10-10-15 bi pawn_imbalance diff
LLR: -3.46 (-2.94,2.94) [0.00,5.00]
Total: 15976 W: 2923 L: 3001 D: 10052
sprt @ 15+0.05 th 1 Try 2, just minor imbalance
10-10-15 bi pawn_imbalance diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 6933 W: 1225 L: 1321 D: 4387
sprt @ 15+0.05 th 1 Bonus for imbalanced pieces and pawn advantage
08-10-15 sn check diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 84377 W: 15693 L: 15453 D: 53231
sprt @ 15+0.05 th 1 Add king support for queen contact checks. The difference should probably be visible only in late queen endgames, so it's difficult to test this in fishtest.
10-10-15 mc tune_imbalance diff
LLR: 2.94 (-2.94,2.94) [0.00,4.00]
Total: 27667 W: 5247 L: 4994 D: 17426
sprt @ 15+0.05 th 1 Test iimbalance tuning after 75K games
08-10-15 mc tune_imbalance diff
37400/50000 iterations
75915/100000 games played
100000 @ 60+0.05 th 1 Tune imbalance at LTC (recently this was a good approach by Stefano)
09-10-15 II lazy_smp_log2 diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 23041 W: 3903 L: 3983 D: 15155
sprt @ 15+0.05 th 3 Additional depth 0-1-3. If this fails, I'm convinced that we have almost perfect formula, and I hope that work on lazy_smp will be continued.
09-10-15 Vo LMRt diff
LLR: -3.08 (-2.94,2.94) [0.00,5.00]
Total: 41454 W: 7711 L: 7662 D: 26081
sprt @ 15+0.05 th 1 LMR tweak
09-10-15 sg attack_scaling diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 8140 W: 1472 L: 1563 D: 5105
sprt @ 15+0.05 th 1 endgame scaling by attacks of stringer side
09-10-15 My CR diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 40838 W: 7652 L: 7600 D: 25586
sprt @ 15+0.05 th 1 Ranked checks?
07-10-15 My Ck diff
LLR: -4.05 (-2.94,2.94) [0.00,4.00]
Total: 74348 W: 13859 L: 13816 D: 46673
sprt @ 15+0.05 th 1 Bigger first check bonus with smaller subsequent checks.
09-10-15 II lazy_smp_log2 diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 12018 W: 1992 L: 2108 D: 7918
sprt @ 15+0.05 th 3 Additional depths 0-1-2. Formula suggested by alwey.
09-10-15 jo lazy_smp_time diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 8645 W: 1400 L: 1489 D: 5756
sprt @ 15+0.05 th 3 Allocate 10% less available time for lazy_smp. This should improve time-management, and also make easy moves more helpful (and visible). (Test with new log formula included, also latest commits from master.)
09-10-15 My CR2 diff
LLR: -2.04 (-2.94,2.94) [0.00,5.00]
Total: 5271 W: 937 L: 1001 D: 3333
sprt @ 15+0.05 th 1 Ranked checks direct from score
08-10-15 II lazy_smp_log diff
LLR: 2.95 (-2.94,2.94) [0.00,5.00]
Total: 25653 W: 3730 L: 3521 D: 18402
sprt @ 60+0.05 th 3 lazy_smp_log vs lazy_smp (3 threads, LTC)
09-10-15 Vo FLT2 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13064 W: 2421 L: 2490 D: 8153
sprt @ 15+0.05 th 1 Fail Low Bonus Tweak 2.
08-10-15 Fi lazy_sharedCM diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 6992 W: 1143 L: 1238 D: 4611
sprt @ 15+0.05 th 3 Share countermoves across threads again
08-10-15 Fi lazy_sharedH diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 774 W: 86 L: 208 D: 480
sprt @ 15+0.05 th 3 Share history across threads again (7523382 is the correct bench not what the commit says)
08-10-15 aj lazy_smp_cmh diff
LLR: -2.94 (-2.94,2.94) [0.00,5.00]
Total: 10816 W: 1683 L: 1763 D: 7370
sprt @ 15+0.05 th 7 Moving CMH to thread: STC
08-10-15 sn bishop4 diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 13763 W: 2518 L: 2584 D: 8661
sprt @ 15+0.05 th 1 Stochastic bishop mobility
07-10-15 II lazy_smp_log diff
ELO: 0.77 +-5.3 (95%) LOS: 61.1%
Total: 4970 W: 765 L: 754 D: 3451
10000 @ 15+0.05 th 23 Lazy_smp - testing log formula, 23 threads.
08-10-15 II lazy_smp_log diff
ELO: 6.79 +-3.4 (95%) LOS: 100.0%
Total: 13716 W: 2435 L: 2167 D: 9114
20000 @ 15+0.05 th 3 Let us also check how lazy_smp_log performs against lazy_smp on 3 threads.
07-10-15 sg lazy_smp diff
ELO: 10.09 +-4.0 (95%) LOS: 100.0%
Total: 9024 W: 1552 L: 1290 D: 6182
10000 @ 15+0.05 th 7 Quick check of lazy smp under hash pressure (see my forum's post). hash=8MB, threads=7. Sorry wrong bench
07-10-15 Vo FLT diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 19629 W: 3581 L: 3622 D: 12426
sprt @ 15+0.05 th 1 Fail Low Bonus Tweak (Fix Bench)
08-10-15 Fi lazy_sharedHCM diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 642 W: 59 L: 180 D: 403
sprt @ 15+0.05 th 3 Share history and countermoves across threads again. Originaly suggested by ajithcj and mbootsector I believe.
07-10-15 mc lazy_smp diff
LLR: 1.13 (-2.94,2.94) [-3.00,1.00]
Total: 28991 W: 4785 L: 4772 D: 19434
sprt @ 15+0.05 th 3 Regression test against original lazy_smp
08-10-15 aj lazy_smp_share_history diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 657 W: 58 L: 178 D: 421
sprt @ 15+0.05 th 7 Use shared history/countermoves for lazy_smp similar to current master: STC
07-10-15 Fi Tempo diff
LLR: -3.08 (-2.94,2.94) [0.00,4.00]
Total: 14126 W: 2083 L: 2199 D: 9844
sprt @ 60+0.05 th 1 Based on SF eval of root position +18cp(46 internal) at depth 42 it seems tempo should be much higher. LTC
07-10-15 aj smp_hybrid5 diff
ELO: -9.24 +-8.0 (95%) LOS: 1.2%
Total: 2256 W: 323 L: 383 D: 1550
5000 @ 15+0.05 th 3 hybrid smp : STC 3 threads low throughput
07-10-15 II lazy_smp_log2 diff
ELO: -22.13 +-15.4 (95%) LOS: 0.2%
Total: 613 W: 77 L: 116 D: 420
10000 @ 15+0.05 th 23 Lazy_smp with the second log formula.
06-10-15 Fi Tempo diff
LLR: 3.10 (-2.94,2.94) [0.00,4.00]
Total: 176331 W: 33104 L: 32325 D: 110902
sprt @ 15+0.05 th 1 Based on SF eval of root position +18cp(46 internal) at depth 42 it seems tempo should be much higher.
07-10-15 mc lazy_smp diff
LLR: 0.19 (-2.94,2.94) [-3.00,1.00]
Total: 32878 W: 5500 L: 5543 D: 21835
sprt @ 15+0.05 th 3 Test for no regression locking semplification
07-10-15 Ro SkipWhenPawnThreats diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 5664 W: 997 L: 1098 D: 3569
sprt @ 15+0.05 th 1 Take 1
06-10-15 aj smp_hybrid3 diff
LLR: 1.30 (-2.94,2.94) [0.00,5.00]
Total: 4679 W: 725 L: 655 D: 3299
sprt @ 15+0.05 th 7 Hybrid of lazy and ybw search: STC
07-10-15 My RCC diff
LLR: 2.96 (-2.94,2.94) [-3.00,1.00]
Total: 15261 W: 2343 L: 2214 D: 10704
sprt @ 60+0.05 th 1 Re-run of a passed simplification to ensure no surprises after several evaluation patches as suggested here (but instead using stricter bounds rather than [0, 4]) https://github.com/official-stockfish/Stockfish/pull/442
06-10-15 Fi lazy_smp diff
LLR: -2.95 (-2.94,2.94) [-3.00,1.00]
Total: 23568 W: 3859 L: 4051 D: 15658
sprt @ 15+0.05 th 3 It should be better to let all threads increment BestMoveChanges. Also simplification. (Testing against fixed easy move.)
06-10-15 Vo HR diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 29135 W: 5368 L: 5368 D: 18399
sprt @ 15+0.05 th 1 Take 3: Update just Fail Low refutation bonus for history.
05-10-15 jh lazy_smp2 diff
ELO: -4.13 +-3.5 (95%) LOS: 1.1%
Total: 10000 W: 1296 L: 1415 D: 7289
10000 @ 60+0.10 th 3 3 thread LTC test for mbootsector's Lazy SMP 2.
06-10-15 Fi Tempo diff
LLR: -2.96 (-2.94,2.94) [0.00,4.00]
Total: 21797 W: 3998 L: 4080 D: 13719
sprt @ 15+0.05 th 1 Based on SF eval of root position +18cp(46 internal) at depth 42 it seems tempo should be much higher. Take 2 using 22. (Take 1 priority dropped to -1)
06-10-15 Vo HR diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 7117 W: 1243 L: 1338 D: 4536
sprt @ 15+0.05 th 1 Take 2: Update just Fail High refutation bonus for history.
06-10-15 aj smp_hybrid2 diff
Pending...
sprt @ 15+0.05 th 7 Second attempt at lazy smp + ybw smp hybrid: STC
06-10-15 Fi lazy_smp diff
ELO: -3.51 +-3.9 (95%) LOS: 4.0%
Total: 10000 W: 1610 L: 1711 D: 6679
10000 @ 15+0.05 th 3 latest lazy_smp w/ easy move fix
06-10-15 aj smp_hybrid diff
Pending...
sprt @ 15+0.05 th 7 a first cut attempt to make a hybrid between traditional and lazy smp: STC
06-10-15 Vo HR diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 28184 W: 5088 L: 5093 D: 18003
sprt @ 15+0.05 th 1 Update History for Fail Low/High Refutation bonus.
06-10-15 aj widen diff
LLR: -2.97 (-2.94,2.94) [0.00,5.00]
Total: 4366 W: 728 L: 835 D: 2803
sprt @ 15+0.05 th 2 Try to widen search at 2 threads: STC
06-10-15 Vo AgeStatRevisit diff
LLR: -2.95 (-2.94,2.94) [0.00,5.00]
Total: 4943 W: 850 L: 954 D: 3139
sprt @ 15+0.05 th 1 Try 50% decay...since we are now dealing with bigger numbers.
06-10-15 Ro JustHanging diff
LLR: -2.96 (-2.94,2.94) [0.00,5.00]
Total: 11525 W: 2107 L: 2183 D: 7235
sprt @ 15+0.05 th 1 Fixed some duplication