Description of the project on the ChessBrain site (excerpts):
ChessBrain is a distributed computation project that uses the idle processing power of distributed machines to solve computationally intensive problems. ChessBrain is a system focused on playing chess, but the underlying system can be adapted for other games as well as for non-game-related applications.
Imagine playing a game against an opponent, except every time he moves, you grab the phone and start calling friends for help. You call Sue, describe the current position and ask her to call you back when she has an answer. Then you call Ryan to ask whether you should worry about a pending attack; again, you ask for a call back when he has an answer. After calling 20 other friends, you sit back and wait for replies. This is similar to how ChessBrain plays chess.
ChessBrain consists of a Linux-based server application, the SuperNode, and client software known as PeerNodes. While ChessBrain plays, it examines positions, dispatches hundreds of potential moves to remote PeerNodes for analysis, collects feedback from the PeerNodes, processes that information and makes its best move.
What about the chess engine? The head of the project, Carlos Justiniano writes: "I reviewed a few, looking for one with relatively clean code and the ability to compile under several operating systems. I found an ideal program in Beowulf, written by Colin Frayn, who was then a PhD candidate at Cambridge University in England. Colin adapted his original Beowulf chess engine to become two chess-playing components, BeoServer and BeoClient. He developed the pair to support distributed chess play within the ChessBrain framework."
On January 30, 2004 ChessBrain attempted a Guinness world record event. Danish GM Peter Heine Nielsen played against a network of 2070 computers. People from over 56 countries participated, the top countries being the US, Denmark, UK, Switzerland, and Germany. Several universities contributed clusters of computers during the event, one university in Denmark using a 250 processor cluster!
The game against Nielsen lasted several hours. The result was a draw by repetition. Peter Heine Nielsen commented that he had set several traps for ChessBrain which computers normally fall for, but was surprised that ChessBrain refused them. Here is the game:
Nielsen,P - ChessBrain [E94]
Guinness record attempt, 30.01.2004
1.d4 g6 2.c4 Bg7 3.e4 d6 4.Nc3 Nf6 5.Nf3 0-0 6.Be2 e5 7.0-0 a5 8.Re1 exd4 9.Nxd4 Bd7 10.Bg5 Nc6 11.Nxc6 Bxc6 12.f3 Qd7 13.Qd2 Rfe8 14.Rac1 h5 15.Kh1 Nh7 16.Bh6 Bxh6 17.Qxh6 Re5 18.Nd5 Rae8 19.Qd2 b6 20.Bd3 Qd8 21.Rf1 Nf6 22.b3 Bb7 23.Qc2 Nd7 24.f4 R5e6 25.e5 c6
The Danish GM has maintained a positional advantage. Now one option would have been to retract the knight and keep up the pressure on the kingside, e.g. 26.Nc3 Nc5 27.f5 Rxe5 28.fxg6 fxg6 29.Bxg6 R8e7 30.Qf2 Rg7 31.Rce1 Rxe1 32.Bf7+ Kh8 33.Qxe1 and the heat is on. Instead Nielsen went for a line that forces a draw by repetition: 26.f5 gxf5 27.Bxf5 cxd5 28.Bxe6 Rxe6 29.Rxf7 Kxf7 30.Qh7+ Ke8 31.Qxh5+ Ke7 32.Qg5+ Ke8 33.Qh5+ Ke7 34.Qh7+ ½-½.