Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!cbmvax!daveh From: daveh@cbmvax.UUCP (Dave Haynie) Newsgroups: comp.sys.mac Subject: Re: 68000 vs 68020 Message-ID: <7260@cbmvax.UUCP> Date: 11 Jul 89 16:21:10 GMT References: Distribution: comp Organization: Commodore Technology, West Chester, PA Lines: 44 in article , sarrel@sioux.cis.ohio-state.edu (Marc Sarrel) says: > A friend of mine just bought at 16MHz 68000 accelerator card for his > SE. How much performance difference is there between a 16MHz 68000 > and a 16MHz 68020? It depends on the system. If you're accelerating a 16 bit machine and not adding any faster memory, a 68020 will in some cases go about maybe 15%-30% faster than the 68000, depending on how well the instruction cache gets utilized. It's actually possible for a 68020 to go slower than a 68000 when on a 16 bit bus, since the 68020 always prefetches 32 bits for instruction processing, though sometimes only 16 bits worth actually gets used. Once you add some 32 bit memory to a 68020, you get real performance increases. The 16MHz 68020 plus some reasonably fast 32 bit memory will get you about 200%-400% speedup over your 7.8Mhz 68000, and a little better than twice the performance of a 16MHz 68000 with fast 16 bit memory. That's all integer performance. If you add a 68881 or 68882 to the 68020, and your software knows about math chips, you may get a floating point performance increase from 25x-100x what a plain 68000 is capable of. I don't know how well the Mac software would integrate any of this, but the hardware driven increases tend to be similar among all 680x0 machines. > Are 68000 accelerators a cost effective alternative? It depends. Without faster 16 bit memory, it's going to boost your performance by maybe 15% or so overall. With the faster memory, it'll come close to doubling the performance, providing the Mac can be set up to use that fast memory before its on-board slower memory. > Can you get faster then 16MHz? Not likely. The 16MHz (actually, 16.6667MHz; Motorola CPUs are rated by the time of the clock cycle, in this case 60ns) part is currently the fastest 68000 anyone makes. > --marc -- Dave Haynie Commodore-Amiga (Systems Engineering) "The Crew That Never Rests" {uunet|pyramid|rutgers}!cbmvax!daveh PLINK: D-DAVE H BIX: hazy Be careful what you wish for -- you just might get it