Below Sent from my iPad
> On Jul 11, 2015, at 8:42 AM, Rhialto <[email protected]> wrote: > > I recall reading that the whole RISC thing was more or less started with > compiler writes for the 360 noticin that in practice they only used a > rather limited subset of the available instructions Not quite true. What started was a paper by an IBM fellow (John Cox). What John noted was that most of the 360/370 systems used a micro architecture below the basic 360 ISA. This micro architecture was always simpler. By this time the assembler vs HLL debate was pretty much over and people were writing in at least systems programming languages (C, PL/360, BLISS much less fortran PL/1 and the like). John asked since we were using optimizing compilers would it not make more sense for the compiler to just target the microengine directly. Patterson looked at Cox's work at took the next step. Let get rid of microengine and reduce the basic instruction set. The term RISC was coined in Dave's paper and the debate was started 😅 Clem _______________________________________________ Simh mailing list [email protected] http://mailman.trailing-edge.com/mailman/listinfo/simh
