Ruversi

a reversi game program written in rust.

Lisence

CC0. in other words, Public domain.

Warranty

NO Warranty.

Command options

• --play
  play a game agaist you. turn is random.
• --playb
  play a game agaist you. your turn is black(SENTE).
• --playw
  play a game agaist you. your turn is white(GOTE).
• --Edax
  play against Edax instead of you. please use with --play(bw).
• --duel N
  play games from some situations with evaltable1 and 2. N is optional. default 5.
• --gtp
  go text protocol mode.
• --ev1 \
  a file for board evaluation.
• --ev2 \
  a file for board evaluation.
• --genkifu
  set generating kifu mode.
• -Nx
  initial board group x for generating kifu. 0~9. all of the initial board positions will be used when this option is not specified.
• --thinkab
  use alpha-beta pruning. default.
• --thinkall
  search every node. (no pruning)
• --learn
  [deprecated] set lerning mode.
• --repeat \
  [deprecated] number of learning. default 10000.
• --eta \
  [deprecated] learning ratio. default 0.0001.
• --rfen \
  think from rfen for debug. don't forget putting "" to an RFEN not to be recognized as 2 part.
• --help
  show help.

Compile options(features)

• nosimd
  stop using simd instructions(SSE) for evaluation.
  SSE floating-point SIMD instructions will be used when no features are specified.
• nnv1
  evaluate with neural network version 1.
• nnv2
  evaluate with neural network version 2.
• nnv3
  evaluate with neural network version 3.
• nnv4
  evaluate with neural network version 4.
• (reserved)nnv5
• avx
  use AVX floating-point SIMD instructions.
• bitboard (default)
  use bitboard instead of byteboard.