Skip to main content
Home Forums 68kMLA Apple IIe Trade - IIGS — #26
Post #26 by Gorgonops
Source Forum68kMLA
CategoryPeripherals
Post DateTue, 16 Jun 2015 - 18:54
Original URLhttps://68kmla.org/bb/threads/apple-iie-trade-iigs.2134/
Post
In this the 65C02 was way ahead of the 6502. And the answer is simple - Being a CMOS processor, the 65C02 has a thinner silicon matrix and less resistance with that matrix. This gives is a faster processing speed on the nano-second level. Given time, nano-seconds adds up to microseconds and microseconds adds up to full seconds. But in today's world, who is going to worry about a couple nano-seconds? Back in the 1980s it was a necessity.
Digital computers derive their timing from a clock; every operation happens according to that cadence so small "analog" differences in the conductivity/responsiveness/whatever between components have *zero* impact on the speed of the machine, everything else being equal. Said differences certainly can "matter"; after all, the difference between, say, 200ns and 150ns-rated RAM is the latter is physically able to respond within a smaller time window, thereby allowing it to be clocked faster, but *if* you have a machine that's designed to use 200ns RAM it's not going to run *ANY* faster with 150ns RAM installed unless you change the memory control circuitry to use faster timing. The same applies to 6502 vs. 65c02 CPUs: if an instruction takes 6 clock cycles on both CPUs it's going to take 6 ticks of the 1Mhz clock, period.

Perhaps the code you were working with happened to use instructions that used fewer clocks on the 65c02, it's certainly possible, but *that* would be the source of the difference, not "less resistance in the silicon".

(The one case where an analog difference might matter is in the case of a peripheral chip that has a "data ready" line the CPU polls and loops on until said data is ready; if the silicon process lets said peripheral stabilize more quickly than the code loop would complete faster, but that's waiting on a *peripheral*. Pretty sure there's nothing *inside* the 65c02 that would apply to.)

mp.ls