grunwald (10/22/82)
#R:populi:-38400:uiucdcs:23900011:000:936 uiucdcs!grunwald Oct 21 18:31:00 1982 Well, I was partially wrong. You can use a 'CMPA' instruction to test an address register, but that requires a total of 12 bytes to test a register for NIL. However, I don't think that that sways the argument very much. I feel that the people at Motorola did a fantastic job with the 68000, but I don't think that it is the end-all of micro-processors. As someone else noted (either here or in net.micro), a lot of bells and whistles don't really make for a fast machine. According to the Aug. Computer magizine, some people who had not done IC design before developed a RISC (reduced instruction-set computer) which would seem to have very fast performance. It only had about 30 instructions, but when one thinks about it, how many does that typical compiler use? I would much perfer that the instruction set be totally orthogonal and not have those little gems like the 'MOVEP' instruction that the 68K has. dirk grunwald