In the mid-1950's, a group of Rice professors wanted a computer to simulate fluid flow. They didn't want to travel to Los Alamos to use the powerful MANIAC II. So, they built their own. The "The Rice Institute Computer", later the R1, began operation in 1959 and was fully functional by 1961. The R1 was known for many things, including a unique (for the time) memory addressing feature known as "codewords", which has been compared to the more modern "object-oriented" view of program design. However, what I found most curious was the "grind crank" single step feature that was added somewhere along the way. Single stepping is a venerable method of program debugging, but if you ever have had to step through large programs, or short ones with long loops, you will understand why a 'Gatling gun' approach would be beneficial. The grind crank would step through 60 instructions per revolution after which you could return to single stepping.
Read all about the history of the R1 here: A Brief History of the Rice Computer, 1959-1971. (the attached photo is from this article) As an aside, I can't dispute the offical history, but I would have sworn that the R1 was still running when I started there in 1972.