Just after the 704 arrived at Livermore (1956?), their 701 went to Corey hall at UC Berkeley. IBM replaced the Williams tubes with a 4Kword core memory just as in the new 704s. That made it into a much more reliable machine. It was nearly twice as fast because memory refresh cycles were unnecessary. Ted Ross wrote a 704 simulator that ran on the 701. It had virtual memory which was swapped out to the 701’s drums, page by page. This was a very early instance of demand paging, in software. Ted also wrote an assembler for the 701 that included some of the superior features of the 704 assembler. What had gone before was very crude. I added some very minor features to the assembler.