z80 timing... 6502 timing

Because of unbelieveably foolish, inane, silly comments like the one you just made, the investigative nature of this problem has been pushed aside if favor of what's really just a game. You've already admitted that you don't even have the totally irrelevant processor you initially claimed you'd use, not that it matters, since that comparison is of no value or interest with respect to the problem at hand. Just read the subject of the email you've been answering. I'm not interested in getting into the equivalent of a virtual computer "tractor pull." I can leave that to the rednecks with the room-temperature IQ's without feeling any loss. However, since the late 1970's there's been this nagging question about the relative merits of two essentially opposed approaches to computer architecture, and a serious attempt to make a comparison without the underlying motive of trying to sell hardware or software has really never been made. I suppose it's because it really doesn't profit anyone to make this comparison at this juncture, except perhaps the incipient 12-year-old lurking within us all. Your comparison would be accurate, perhaps, if it took only a couple of days' effort and parts we all probably have lying about anyway to build that formula-1 car you refer to, but if a person wants to perform a valid test, there has got to be suitable hardware. Now, there are lots of plain-vanilla-flavored microcomputers with a Z-80 at their hear. However, almost all the 6502-based computers were really designed for the video-toy market and only find themselves able to serve as computer only as an afterthought. The Apple-II is probably a prime example. It would be acceptable to run a comparative exercise on the two processors if some valid basis for the comparison could be determined. Hans Franke came up with a proposed means for evaluating the validity of each arrangement. Unfortunately, this requires a test system be prepared, with a suite of test software and some specialized hardware. It also requires that each system-under-test be equipped with compatible hardware. I find this proposal valid, but quite a way off the mark, in that it makes the test almost completely hardware dependent. My take on this test would be to permit development and execution of the algorithm ultimately deemed most appropriate for this test on whatever hardware the programmer has at his/her disposal. It should be as limited in its hardware requirements as possible, i.e. it should not matter whether the program is written for a TIMEX Sinclair, or a CRAY MPX with simulation capability. The sum total of the resources involved in the code limited to an amount of memory common to all the contestant systems, and a console interface common to all as well. Code for interacting with the system console need not be considered, so long as it is entered with a call to a routine requiring NO preparation prior to the call. That means you must call a routine to make the console I/O preparations before calling the console I/O routine itself. The time for the first call and return made within the contestant's code is to be included in the competition, but subsequent action is not. However, the called console I/O code must be provided in order to show that no task-related effort is being made by the console handling routine. This is simple with a terminal, but not so simple with a device having resident video and keyboard I/O. Hans Franke suggested this test be performed on a system like the KIM-1. It is probably achievable in one. It must be so limited that no one system can outperform another just because of its resources. The comparison is between the processors, not the systems in which they reside. The code applicable to the contest can be evaluated for its consumption of processor cycles and relative timing computed from that. However, a lowest common denominator with respect to resources must be applied in the strictest sense. All of this monitoring and calculation can be eliminated, however, if each contestant simply builds a simple system with a full compliment (whatever that is determined to be) of memory, and both of the subject processors can use only 64K of memory. The resources implicit in the processor design shouldn't be an issue as that's part of what's being compared. Consequently it must be inherently permissible to use as much stack as needed without being penalized in any way, provided that falls within the designated limits. Again, such limitations go away if everybody runs their processor in 64K of RAM/ROM with only the one I/O device. Doesn't this make more sense than having a "tractor-pull" between computers? Dick -----Original Message----- From: Sellam Ismail <dastar(a)ncal.verio.com> To: Discussion re-collecting of classic computers <classiccmp(a)u.washington.edu> Date: Sunday, April 18, 1999 12:10 AM Subject: Re: z80 timing... 6502 timing