As with all computers of its era, it was a one-of-a-kind machine. The name was chosen in the hope of stopping the rash of silly acronyms for machine names. MANIAC weighed about 1,000 pounds.
The first task assigned to the Los Alamos MANIAC was to perform more precise and extensive calculations of the thermonuclear process.
The lab got hold of one of the computer, one of the first, so that its scientists, H-bomb researchers, wouldn't have to stay up nights solving their voluminous equations with pencil and paper. In 1953, the MANIAC obtained the first equation of state calculated by modified Monte Carlo integration over configuration space.
Remember the Monte Carlo analysis feature in Rybka that was so eagerly hyped several years back? Basically with the Monte Carlo analysis you start with a position, activate the feature, and then the Rybka engine played a whole lot of games (hundreds or even thousands) against itself until you stopped it and then it stored all of the moves of those games in tree form and you could see the results of games played from the position under investigation.
The MANIAC ran from March, 1952 until it was shut down on July 15, 1958. However, it was transferred to the University of New Mexico in bad condition and was restored to full operation by Dale Sparks, PhD. It was featured in at least two UNM Maniac programming dissertations from 1963 and remained in operation until it was retired in 1965. It was succeeded by MANIAC II in 1957. A third version MANIAC III was built at the Institute for Computer Research at the University of Chicago in 1964.
In 1956 mathematician and mathematical physicist Professor Stan Ulam had done some very important work in connection with atomic energy involving “computing machines”, as they were known in those days, and he and his colleagues had also written about chess playing machines.
Normally, Professor Ulam, a decent chessplayer himself, worked at the Los Alamos (New Mexico) Scientific Laboratory, but later was a visiting professor in the mathematics department at MIT where he conducted experiments in which a computer was able to play a game on the basis of operating instructions of only a very general nature.
Earlier, Claude Shannon had succeeded in getting a small homemade computer to play a few simple position without exceeding the computer’s memory.
In a letter to the editor from Edward Lasker that appeared the January, 1957 issue of Chess Review, at the request of Claude Shannon, Lasker sent them an article by Stan Ulam and his colleagues that discussed a chess program they had written for MANIAC I.
Lasker stated that he had discussed the subject of chess-playing computers with a few players who were not familiar with Shannon’s paper and they flatly refused to believe that a machine could play chess and the idea of a machine “thinking” was “violently” rejected!
MANIAC I could perform 10,000 elementary operations per second. It selected its moves based on two factors, the greatest mobility and material. The very first game pitted the computer against itself. The first human that lost to the computer was an unidentified lady who had learned the game only a week earlier; it was the first time a human had lost to a computer in a game of intellectual skill.
Early on, moves averaged about 10 minutes for a two-move look ahead, but that quickly became three moves, four moves, five moves ahead and it quickly reached the capability of seeing 12 moves ahead.
At one time when Princeton physicist Martin Kruskal checkmated the MANIAC, the machine responded with an illegal move. Programmers traced the problem to the fact that the computer had not been programmed to resign and as a result got stuck in a loop.
No comments:
Post a Comment