Firenzina edges Stockfish, beats IvanHoe playing gambits

[Gusev]

I created a new Firenzina wiki, http://firenzina.wikispaces.com/home, and posted PGNs of three 4/4 50-game matches played by x64 compiles of (a) Firenzina 2.2.2 xTreme, a clone of Fire 2.2 xTreme by Kranium, the MT_SGVAD_LP compile; (b) Stockfish 2.3.1 x64 popcnt_ja (CCRL 40/4 3215), and (c) IvanHoe 999946h, the KLO compile (CCRL 40/4 3080). Conditions of the internal test were as follows.
- opening book: silvergambit.pgn
- The number of threads was set to 8 for each engine.
- Ponder off.
- Large pages were enabled for Firenzina and IvanHoe.
- RobboBases were enabled and configured for Firenzina and IvanHoe.
Results:
- Firenzina 2.2.2 xTreme beat IvanHoe 999946h, +19-3=28 (33:17, 66%)
- Stockfish 2.3.1 (Stockfish-231-64-popcnt-ja) was victorious over IvanHoe 999946h, +15-10=25 (27.5:22.5, 55%)
- Firenzina 2.2.2 xTreme was lucky to edge Stockfish 2.3.1, +13-10=27 (26.5:23.5, 53%)
These results are posted here not to make any definite claims about relative strengths of chess engines, but merely to encourage others to test Firenzina some more. In particular, it is well understood that 4/4 is not the same thing as 40/4 calibrated to AthlonX2. The results make some sense after Firenzina's February compile got 56.2% against Stockfish 2.3.1 in marijan's 1 min gauntlet rating list, http://immortalchess.net/forum/showthread.php?p=348400, having also scored 52.8% and 58.3% against two other Stockfish versions. Basically, Firenzina is seemingly hanging somewhere near Stockfish, and more testing on longer time controls would do good.