I don't really care what's done here, but I thought the object was to see
which of the two processors in the "subject" field would make for faster
execution of a useful and published algorithm than the other.
It's necessary that the algorithm be implementable on similarly equipped
platforms, whether they exist physically or not. The limitation is that it
must be timed as executed on a device from one of the two processor types in
question, and physically available in 1982-83 as that's the time-frame about
which the discussion preceding this coding exercise was centered.
There's no point in specifying it for an Apple-II, because that one didn't
even run the processor at the current maximum rate due to its overlap with
video display refresh timing, and because it would involve too much
unrelated design and construction effort to come up with a suitable
substitute implemented using a Z-80. Consequently I proposed one wire his
own computer using the processor, 64K of static ram, and a serial port of
some type TBD. Maybe, just for the exercise, a file device, e.g. a floppy
disk controller ala WD1770/72 ought to be included. That's got to be hashed
out for sure, if it's to be realized in hardware. Problems potentially lie
in the path, however, as some of the hardware may be scarce if availalble at
all.
A suitable port for attaching a terminal or PC ought to be included, but
only in its most basic form. That way, when the builder is finished, he has
an item he could possibly use for something, should he choose to do so.
This can all get to be a mite burdensome when all you wanted was to see what
the fastest or most efficient code one could come up with would look like,
which I why I suggested a simulator. The only problem with that is that one
could then write and assemble code which didn't in reality do what it
claimed because the I/O wasn't simulated as well.
This requires some more thought.
Dick
-----Original Message-----
From: Sellam Ismail <dastar_at_ncal.verio.com>
To: Discussion re-collecting of classic computers
<classiccmp_at_u.washington.edu>
Date: Saturday, April 17, 1999 2:22 PM
Subject: Re: z80 timing... 6502 timing
>On Sat, 17 Apr 1999, Richard Erlacher wrote:
>
>> Well . . . There's the problem . . . first of all, the code's got to be
>> executable on something everyone has got available, or it's got to be
>> simulated on a simulator everyone has available, else there'll be a limit
on
>> interest right away . . . Then, shouldn't there be some consideration of
>> the coding/debugging time involved? I'd lean in favor of a PC-compatible
>> simulator. That makes the computation of actual execution time
>> straightforward. . .
>
>Not necessarily. You measure the code based on an analysis of the clock
>ticks it uses. This way the competition is platform independent. Of
>course a suitably platform independent code spec would need to be
>developed.
>
>> . Then there's the question about WHICH 6502 to use. Given a listing,
it's
>> easy enough to compute how long it takes the code to run, but which
>> instruction set? What about undocumented features? Both these
processors
>> were famous for those. Of course, there doesn't have to be a limitation,
>> i.e. one could consider ALL available cores.
>
>This would not be limited to the 6502. The idea is to see who can come up
>with the most efficient algorithm on any processor.
>
>Sellam Alternate e-mail:
dastar_at_siconic.com
>---------------------------------------------------------------------------
---
>Don't rub the lamp if you don't want the genie to come out.
>
> Coming this October 2-3: Vintage Computer Festival 3.0!
> See http://www.vintage.org/vcf for details!
> [Last web site update: 04/03/99]
>
Received on Sat Apr 17 1999 - 16:11:11 BST