It might be assumed that it is the double sized net, but I don't think so.
I am getting about 60M NPS, where I used to get about 90M, which would account for less than a ply.
I am seeing lots of very slow searches just to reach 30 ply.
I wonder what will happen to SF in long time control with powerful hardware matches when it is out-searched by 15 plies.
Try searching these with the latest abrok stockfish and a version about a month older:
- Code:
-
1n1qr1k1/r4pp1/pbppBn1p/1p2p3/PP2P2B/2PP1N2/3NQPPP/R4RK1 b - -
2r5/3qb2k/1p2p1rp/pP1pPp2/1n1P1Pn1/1B5R/1P1BQ1PN/5R1K b - -
r1b2k1r/ppp1q1b1/2np1n1p/8/3PPB1P/1QP2B2/PP4P1/RN3RK1 b - -
r1bq1rk1/2p2ppp/p1np1n2/1pb1p1B1/4P3/1BPP1N2/PP2QPPP/RN2K2R b KQ -
r1bq1rk1/ppp2ppp/2np1n2/4p1B1/2B1P3/2PP1N2/P1P2PPP/R2Q1RK1 b - -
r1bqk1nr/ppp2ppp/1bnp4/4p3/P1B1P3/2P2N2/1P1P1PPP/RNBQ1RK1 b kq -
r1k3nb/ppp2p2/2n3q1/3NPbp1/2B2p2/6P1/PPPB3N/R3Q1K1 w - -
r2q1rk1/2p2pp1/p1npbn1p/1pb1p3/4P2B/1BPP1N2/PP1NQPPP/R3K2R b KQ -
r3k1nr/p1pn1pb1/1p1p3p/3P2p1/PPN1PpqP/2P2N2/6P1/R1BQ1RK1 w kq -
r4rk1/p1qn1ppp/4p3/2bn4/2Np4/NQ6/PPP2PPP/R1B1R1K1 b - -
rn1k1bnr/pp1q3p/2pp2b1/6P1/3PPBQ1/2NB4/PPP3P1/R4K1R w - -
rnb1k1nr/pp2qpb1/2pp3p/6p1/2BPPp2/1QP2N2/PP4PP/RNB2RK1 w kq -
rnb1k1nr/pppp1pb1/7p/6pq/2B1Pp1P/2N2N2/PPPP2P1/R1BQ1K1R w kq -
rnb1k1nr/pppp1pbp/8/6pq/2B1Pp2/2N2N2/PPPP2PP/R1BQ1K1R w kq -
rnb1kbnr/ppp2p1p/3p4/6pq/2BPPp2/5N2/PPP3PP/RNBQ1K1R w kq -
rnbq1r1k/ppp3pp/5n2/2bPp3/2B1P3/2N2N2/PP4PP/R1BQK2R w KQ -
rnbqk1nr/ppp2ppp/8/2bP4/4pP2/5N2/PPPP2PP/RNBQKB1R w KQkq -
rnbqk2r/p1p3pp/5n2/1p1Ppp2/2p1P3/2N2P2/PP4PP/R1BQKBR1 b Qkq -
rnbqkb1r/p1p1pppp/5n2/1p1P4/8/1B3P2/PPPPQ1PP/RNB1K1NR b KQkq -
rq3rk1/1pb2ppp/2p1bn2/p2p4/B4P1Q/P1NP3P/1PPBR1P1/R5K1 b - -
Home 
