"It was incredibly slow for an all-electronic device"
It wasn't all-electronic; that memory used _sound_waves_ to store bits. That six meters of wire probably meant that reading a word, worst-case, took about a ms (assuming a fairly generous sound speed of 6km/s), and about every instruction read from and wrote to that storage. This being the first of its kind, the implementation probably would be worst-case for both reading and writing (wait for the 'start of memory' marker, count bits until the ones needed come along)
And of course, it had to read its instructions from the same wire. I guess 500 instructions per second would be an exaggeration of its performance.
I started wondering how one would reliably store a bit in about a microsecond of audio in the 1960s in a device that doesn't require continuous tuning.
It wasn't all-electronic; that memory used _sound_waves_ to store bits. That six meters of wire probably meant that reading a word, worst-case, took about a ms (assuming a fairly generous sound speed of 6km/s), and about every instruction read from and wrote to that storage. This being the first of its kind, the implementation probably would be worst-case for both reading and writing (wait for the 'start of memory' marker, count bits until the ones needed come along)
And of course, it had to read its instructions from the same wire. I guess 500 instructions per second would be an exaggeration of its performance.