16 Apr
1999
16 Apr
'99
3:50 p.m.
Since we all see to have so much time on our hands . . . perhaps it would be
amusing to attempt to write the SAME logic in the SAME way on each
processor, then compare to see which is faster. THEN turn around and try to
improve the implementation of this logic to see which is easier to speed up,
and perhaps see which has the greatest cost in terms of memory usage, etc.
It's truly a retrocomputing exercise, as neither processor is "still
around"
in the sense of mainstream applications, or even as an "active" product.
I guess that qualifies it as an extended topic in this forum.
Dick
-----Original Message-----
From: allisonp(a)world.std.com <allisonp(a)world.std.com>
To: Discussion re-collecting of classic computers
<classiccmp(a)u.washington.edu>
Date: Friday, April 16, 1999 2:00 PM
Subject: Re: z80 timing... 6502 timing
I realy love
your mails - Just reading all of this I felt almost
20 years younger - I was involved _very_ heavyly into the Z80/8080
vs. 6502 discusion back then - weh had a circle of engeneers where
one part wher 8085/Z80 advocates, whie some (like me) where 6502
boreheads.
Actually I'm not done with it. just have the time. The fundemental
problem is not really which is faster, I think for the real world
applications they are likely close but the arguement is flawed. It
misstates the z80 and 6502 timing vs real useage.
What's funny is I have a kim one I can run and a z80 SBC for comparison.
Allison
16 Apr
16 Apr
10:57 p.m.
On Fri, 16 Apr 1999, Richard Erlacher wrote:
Since we all see to have so much time on our hands . .
. perhaps it would be
amusing to attempt to write the SAME logic in the SAME way on each
processor, then compare to see which is faster. THEN turn around and try to
improve the implementation of this logic to see which is easier to speed up,
and perhaps see which has the greatest cost in terms of memory usage, etc.
It's truly a retrocomputing exercise, as neither processor is "still
around"
in the sense of mainstream applications, or even as an "active" product.
Ooh! A machine code competition. I'm in! I'll do the 6502 and whoop
EVERYONE'S ass!
Sellam Alternate e-mail: dastar(a)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]
17 Apr
17 Apr
3:04 a.m.
Sellam wrote:
Ooh! A machine code competition. I'm in!
I'll do the 6502 and whoop
EVERYONE'S ass!
Sounds like I challenge; I'm a damn good 6502 programmer, IMNSHO.
On the other hand, I've found that the 6809 often has better code density
than any other 8-bit micro. Many years ago there was a challenge on one of
the newsgroups to write the smallest routine (not necessarily the fastest)
to convert a number from 1-255 in the accumulator to roman numerals (output
by a subroutine call that prints a character). I don't remember the byte
counts, but my 6809 routine was several bytes shorter than any other
entrant. There was one famous net.personality that insisted that his 6502
code was even shorter, but he refused to post the code so his claim went
unverified.
To make my code as short as it was, I used every trick in the book and
then some. For instance, I had a table whose end was delimited by a byte
with the high bit set. The table was placed immediately *before* the code
in memory, because the high bit of the first opcode was set and could thus
end the table.
3:18 a.m.
Ooh! A machine code competition. I'm in!
I'll do the 6502 and whoop
EVERYONE'S ass!
And then you woke up. First what are the rules, 6502 or 65C02, code in Rom
or Ram, what is the code supposed to do?
My vote goes for something with some graphical element so we can "see" what
is happening. (spinning ball, etc.)
3:37 a.m.
On Sat, 17 Apr 1999, Mike Ford wrote:
BRING IT ON, MO FO!
Sellam Alternate e-mail: dastar(a)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]
Ooh! A machine
code competition. I'm in! I'll do the 6502 and whoop
EVERYONE'S ass!
And then you woke up. First what are the rules, 6502 or 65C02, code in Rom
or Ram, what is the code supposed to do?
My vote goes for something with some graphical element so we can "see" what
is happening. (spinning ball, etc.) 
10:09 a.m.
New subject: Coding chanllange (was: z80 timing... 6502 timing)
> >Ooh! A machine code competition. I'm
in! I'll do the 6502 and whoop
> >EVERYONE'S ass!
> And then you woke up. First what are the rules,
6502 or 65C02, code in Rom
> or Ram, what is the code supposed to do?
Good point.
> My vote goes for something with some graphical
element so we can "see" what
> is happening. (spinning ball, etc.)
Nice, but already to high level.
The task should be abstract enough to run within a simple
system, lets say a SBC system with CPU, some MEM and some
of the usual I/O&Controll stuff (Ports, Serial, Timer).
Basicly there are two types of possible challanges:
a) implements some kind of standard code in a most speedy way, or
b) build some small, but sophisticated "real world" application.
Challenge a) would meet the basic idea we had in our discusion:
define at what ratio two processors are equivalent when executing
a real task, while b) should be sophisticated enough to give a
'boh ey' (sound of exeptional admire) effect.
In terms of a sportive competition a) would be a compulsatory
exercise, while b) is more a voluntary (free style) exercise.
As I stated before, for our theme a) would be more aprobiate.
As actual tasks I would suggest
for a): a 'simple' sorting challenge
given is
- Input media
- Input structure
- input data (same data or all participants, randomly generated)
- Type of sort algo to use (what about just a bubble ? :)
- Output structur
- Output media
the score schould be made up from
Execution time (mainly for the sort, but also for I/O)
Programm length (again most points to be given on the sort itself)
Style (readability, portability, maintainability)
for b) a thing like a small multi tasking OS could be used,
maybe for a controlling application (heating or traffic -
or what about a model railroad setup ?) or other purposes ?
(This could be chalenging, since these kind of tasks are
maybe a bit tough for our little helpers :)
So, you're opinion ?
BRING IT ON, MO FO!
Don't rub the lamp if you don't want the genie to come out.
MO FO ? I fear, I'll have to rub the lamp.
Gruss
H.
--
Stimm gegen SPAM: http://www.politik-digital.de/spam/de/
Vote against SPAM: http://www.politik-digital.de/spam/en/
Votez contre le SPAM: http://www.politik-digital.de/spam/fr/
Ich denke, also bin ich, also gut
HRK
3:55 p.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 1344 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/19990417/31190848/at…
4:11 p.m.
On Sat, 17 Apr 1999, Sean 'Captain Napalm' Conner wrote:
I would vote against that because the hardware
available might make a
difference. For instance, moving an object on a 6502 based system, the C-64
will probably win both the speed and size catagories, simply because of
hardware assist. Besides, if this is to match code/speed size between CPUs,
then not all machines are capable of graphics.
No graphics.
The Roman Numeral one sounds interesting. To make
it fair, I would assume
that there exists a routine, CHROUT, that takes as input the character in
the main accumulator and displays it on the output device. The output
device is a simple TTY like device that supports TAB, CR (which returns the
print head to the start of the line, and advances to the next line) and BS.
Not necessary. The resulting ASCII string could simply be stored at an
address in memory.
Of course the Roman Numeral problem, while a good one, would give Eric an
unfair advantage since he could just go back and optimize his existing
algorithm, and Eric doesn't need a head start. ;)
The code for CHROUT is NOT to be counted towards
speed or size, but I'd
probably allow the call to CHROUT to be considered.
Contest open for any CPU anyone would care to write code for.
-spc (Hows that for a contest?)
A hell of a lot more simple than the other proposals being floated.
Sellam Alternate e-mail: dastar(a)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]
20 Apr
20 Apr
5:02 a.m.
What might speed this along is if we create a list of agreeable items,
instead of speculating in space. As I have been reading the messages a list
is forming for me.
Some things in this contest that sound reasonable to me;
Input and output are to memory resident buffers.
Inline code is too boring, and subrountine calling too important, so the
"task" should require perhaps modules; maybe make the contest half a dozen
subroutines which get called from a contest defined main program (using
some typical non asm language like C or pascal).
Needs to be fun, NOT work, and should provide a good excuse to play with
our old hardware.
Contest submissions could be a simple binary file, file length, predefined
jump table for subroutines, the actual code. Some third party, can run the
code and time it on the hardware of their choice.
7:41 a.m.
Some things in this contest that sound reasonable to
me;
Input and output are to memory resident buffers.
Inline code is too boring, and subrountine calling too
important, so the
"task" should require perhaps modules; maybe make the contest half a dozen
subroutines which get called from a contest defined main program (using
some typical non asm language like C or pascal).
Contest submissions could be a simple binary file,
file length, predefined
jump table for subroutines, the actual code. Some third party, can run the
code and time it on the hardware of their choice.
Nice, but with these things, like internas of the system for input/output,
binary and similarities, you tie again all down to a single system to
use - we loose the idea of a cross platform competition where only basic
processor features are measured (see also the subject).
Gruss
H.
--
Stimm gegen SPAM: http://www.politik-digital.de/spam/de/
Vote against SPAM: http://www.politik-digital.de/spam/en/
Votez contre le SPAM: http://www.politik-digital.de/spam/fr/
Ich denke, also bin ich, also gut
HRK
22 Apr
22 Apr
1:05 a.m.
Some things in
this contest that sound reasonable to me;
Input and output are to memory resident buffers.
Inline code is too boring, and subrountine
calling too important, so the
"task" should require perhaps modules; maybe make the contest half a dozen
subroutines which get called from a contest defined main program (using
some typical non asm language like C or pascal).
Contest submissions could be a simple binary
file, file length, predefined
jump table for subroutines, the actual code. Some third party, can run the
code and time it on the hardware of their choice.
Nice, but with these things, like internas of the system for input/output,
binary and similarities, you tie again all down to a single system to
use - we loose the idea of a cross platform competition where only basic
processor features are measured (see also the subject). 
17 Apr
17 Apr
5:42 p.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 940 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/19990417/e21ed81a/at…
6:18 p.m.
On Sat, 17 Apr 1999, Tony Duell wrote:
Sure, why not! That would be great.
Contest open
for any CPU anyone would care to write code for.
Including one that I've designed? In which case, how general-purpose does
it have to be? I have half an idea to make a state machine that does
binary -> roman conversions, but not anything else. I suspect I could get
it rather fast, though. (Yes, that's cheating. I know it's cheating.
But it's not specifically
disallowed by the above).
I don't think anyone should have to be handicapped. Of course, you'll
have to demonstrate that it actually work, although I have no doubt you
can and will :)
Sellam Alternate e-mail: dastar(a)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]
8:09 p.m.
A non-text attachment was scrubbed...
Name: not available
Type: text
Size: 941 bytes
Desc: not available
Url :
http://www.classiccmp.org/pipermail/cctalk/attachments/19990418/33a759a2/at…
9393
days inactive
9399
days old
13 comments
7 participants
participants (7)
-
ard＠p850ug1.demon.co.uk -
dastar＠ncal.verio.com -
edick＠idcomm.com -
eric＠brouhaha.com -
Hans.Franke＠mch20.sbs.de -
mikeford＠netwiz.net -
spc＠armigeron.com