louie@umd5.UUCP (08/30/85)
In article <229@tikal.UUCP> dal@tikal.UUCP (Don Ledford) writes: >------ >In article <916@vax2.fluke.UUCP> kurt@fluke.UUCP (Kurt Guntheroth) writes: >>The amiga uses the 68000 chip exactly because it is less expensive and the >>amiga does not require the 68010 extra capabilities of instruction restart. >>The 68000 and 68010 are directly compatible and if you want to speed >>everything up about 7-8%, you can just drop in a 68010 (assuming the 68000 >>is in a socket of course). >>-- > >I'm a software type and don't have access to the amiga schematics, but >I would caution against replacing the 68000 chip with a 68010 until >you know for sure it doesn't effect the shared access to the video RAM. >It may be that access is arbitrated by monitoring the instruction >fetch activity of the CPU (during fetch the DMA has a free shot at >the video RAM). The 68010 doesn't need to perform fetches for certain >types of code loops (hence the speed improvement); this may cause >unacceptable delays for the video circuitry. > >Speculation only. > >Don Ledford [I'm moving this discussion to net.micro.amiga] There is a more fundamental problem; the state of the stack during expection processing for a 68010 is not identical to the 68000's. This only is of concern to the underlying operating system and if it was clever enough could probably handle both. I don't remember for sure, but this may only be a problem for bus error type exceptions. Can't wait for my Amiga to get here! -- Louis A. Mamakos WA3YMH University of Maryland, Computer Science Center Internet: louie@trantor.arpa UUCP: {seismo!umcp-cs, ihnp4!rlgvax}!cvl!umd5!louie
jjg@ccice5.UUCP (John Grana) (09/07/85)
> > There is a more fundamental problem; the state of the stack during expection > processing for a 68010 is not identical to the 68000's. This only is of > concern to the underlying operating system and if it was clever enough could > probably handle both. I don't remember for sure, but this may only be a > problem for bus error type exceptions. > > Louis A. Mamakos WA3YMH University of Maryland, Computer Science Center In the case of the 68010, this is a problem for ANY exception. The 68000 places 3 words on the stack (sr,pc), while the 68010 'dumps' 4 ( sr, pc, vector#/format)! john grana C.C.I Rochester, NY jjg@ccice5