The Network Infrastructure for Combinatorial Exploration, or NICE, is specifically designed to support our research on distributed search strategies.

NICE supports both nagging (a parallelization scheme first developed in the context of our work on the DALI theorem prover) and partitioning, the traditional search parallelization scheme. NICE allows application programmers to parallelize their search procedures by simply making appropriate function calls within their code. The distribution includes niced, a Unix resource management daemon, niceq, a daemon status query program, and niceapi, the applications programmer's interface library. The code is written in ANSI C with BSD sockets over TCP/IP, and is known to run on multiple variants of the Unix operating system, including Linux, Sun OS, and HPUX.

Sample NICE applications include a parallel Davis-Putnam algorithm for 3SAT, a Traveling Salesman solver based on a parallel A* search algorithm, and a game-independent parallel alpha-beta minimax algorithm. NICE also serves as the basis for our work on applications to the biosciences, including HOPS and MLIP.