jhs@MITRE-BEDFORD.ARPA.UUCP (04/30/87)
Aha! A good argument is developing... Eric Greene writes: >the fastest 68000 instruction is but 4 cycles long - the average instruction >is around 12-14 cycles. Compare this to the 6502's AVERAGE of about 4-5 >cycles per instruction (only CLC, INY, etc. take 2 cycles). At about 4 times >slower than the 8.0 MHz 68000, NOTHING is faster on a 6502. I hate to be stubborn (actually I LOVE to!) but Eric's own figures tend to support my claim, at least to the extent that there are some tasks for which there wouldn't be a vast difference in speed between a 6502 and a 68000. If we compare AVERAGE times, based on Eric's figures, then the average time per instruction for the two machines would be: 8-MHz 68000 1.76-MHz 6502 ------------------------------------------------------------- 13 cycles (avg) 4.5 cycles (avg) ------------------ = 1.625 usec ------------------ = 2.556 usec 8.0 MHz clock 1.76 MHz clock for a 1.573:1 advantage in favor of the 68000. This is considerably less than the 4.5:1 ratio of the clock speeds or the "4 times faster" that Eric alludes to. Probably much of the power of the 68000 comes from getting more power per instruction rather than more instructions per second. -John Sangster