Re: Mini computer
It isn't just the clock speed, it is also I/O and other things that were optimized for mutli-user and multi-process use.
The VAX 11/780 had a 5Mhz processor and ~4MB RAM, but could cope with over 100 concurrent users. Today, you would be hard put to find a PC that could handle 100 concurrent users. I worked for an oil exploration company in the mid-80s. They had dozens of geologists working on seismic plots and writing FORTRAN code to clean up the raw data, A0 digitising tables for marking areas of interest on printouts etc.
They even had a 2400dpi laser plotter (which used an Olivetti mini-computer as a dedicated print server).
The same for the PROTOS2000 ERP system, running on a VAX. We had that at a manufacturing company with hundreds of users in the back office and on the shop floor entering data or analysing results.
Even considering that was mostly 80x25 or 132x25 character terminal displays, that is still a lot of capture, storage, retrieval and processing of information that it had to do with those 5mHz. It had to have a separate process for each terminal, for example.
Modern computers have much more processing power, but I honestly don't know if they could cope with hundreds of users hanging off of them and using them to process information simultaneously. But getting hold of such an optimised operating system and a method of connecting a couple of hundred serial terminals to a Pi would be near impossible these days (yes, you could probably do 100+ Ethernet based terminal sessions).
Even the first PCs were, theoretically, "faster" than a mini computer of that age, in ram Mhz terms, but they struggled when you tried connecting more than half a dozen terminals to the back of them - I had to support a 286 running Xenix and an accounting package, with a dozen terminal hung off a multi-port serial card. It was very slow, even though its processor was faster than a VAX, in terms or clock speed. What the VAX could achieve in one clock cycle, on the other hand...