Comments on: Maze Generation In Thirteen Bytes

By: Gloriosas inutilidades: gerando um labirinto em 10 (sim, DEZ) bytes! | Retrocomputaria Plus

Mon, 04 Mar 2013 12:26:23 +0000

[...] os recordes vão se sucedendo: 25 bytes, 15,13, 12. 11… dez! A essa altura o povo já estava apelando para truques que não são coisa de Jesus, tais como [...]

By: Trixter

Trixter — Tue, 08 Jan 2013 18:07:11 +0000

I guarantee you it does not work on NEC V20 and V30 CPUs. D6 on those CPUs is a duplicate of XLATB which is D7. More background and tests: http://www.vintage-computer.com/vcforum/showthread.php?28949-The-D6-instruction-on-the-NEC-V-series-(V20-V30)-chips

Good to know it works on everything else. BTW I don’t consider this a “bug” in the NEC chips; it wasn’t documented when they were being designed.

By: Peter Ferrie

Peter Ferrie — Tue, 08 Jan 2013 17:38:15 +0000

Thanks.
It should work on all Intel, AMD, and Cyrix processors.
Intel introduced the instruction, AMD documented the instruction, and Cyrix copied the instruction.

By: Andrey Baranovich

Andrey Baranovich — Tue, 08 Jan 2013 09:52:32 +0000

Cool! :-)

By: Post Position » 10 PRINT Marches on

Post Position » 10 PRINT Marches on — Tue, 08 Jan 2013 04:04:05 +0000

[...] Trixter provides the full story (so far!) on his blog, Oldskooler Ramblings. [...]

By: Trixter

Trixter — Tue, 08 Jan 2013 00:48:36 +0000

The more I look at what you did, the more I am a fan :-) I updated the article with your code.

By: Trixter

Trixter — Tue, 08 Jan 2013 00:36:23 +0000

AMAZIN– hey, wait a minute, that uses an undocumented opcode that doesn’t work on all processors.

So I guess all three of us have our own record — mine for working on any machine, herm1t for working on any machine running DOS, and you for working on any Intel CPU. We all winnerz! \o/

By: Peter Ferrie

Peter Ferrie — Mon, 07 Jan 2013 22:05:29 +0000

“10 Print” in 10 bytes!!

init:scasb
salc
and al,’\'-’/’
add al,’/’
int 29h
jmp init

By: ange albertini

ange albertini — Mon, 07 Jan 2013 11:30:39 +0000

herm1t chopped another byte from the “AH = 0xE + Int 10h” part by using the ‘undocumented’ int 29h.

cf http://lj.rossia.org/users/herm1t/66490.html

By: Bruno

Bruno — Fri, 04 Jan 2013 00:02:22 +0000

I haven’t had this much fun reading a post in a long, long time :)

By: holograph

holograph — Fri, 21 Dec 2012 15:00:04 +0000

Can’t blame you, semantically speaking you’d definitely expect SCAS to use DS:SI… surprising!

By: Trixter

Trixter — Tue, 18 Dec 2012 15:08:27 +0000

I’m Doing It Wrong! Shows you how much I use SCAS in my code. Fixed.

By: Peter Ferrie

Peter Ferrie — Tue, 18 Dec 2012 04:59:23 +0000

Intel docs: “SCAS m8: Compare AL with byte at ES:(E)DI or RDI, then set status flags”.

By: Trixter

Trixter — Tue, 18 Dec 2012 00:25:44 +0000

I added the “where” but I left DS:SI alone — SCASB doesn’t read from ES:DI.

By: Peter Ferrie

Peter Ferrie — Mon, 17 Dec 2012 23:33:36 +0000

“read from DS:SI points” -> “read from where ES:DI points”

By: Trixter

Trixter — Mon, 17 Dec 2012 19:40:49 +0000

Because the code segment is used as the random data source, it does not look great if you boot and then immediately run the program. Run any other program first, preferably a large one, to fill the first 128K of RAM with junk.

By: phoenix

phoenix — Mon, 17 Dec 2012 19:28:34 +0000

Does this work better on a real DOS PC than elsewhere? In a Win XP box, it gives mostly \ chars with flashes of random. In DOSBox, the random flashes are even less frequent.

By: phoenix

phoenix — Mon, 17 Dec 2012 19:11:37 +0000

If there’s anything to be learned about size coding, never ever say “I believe this is as small as it can get”. :)

By: Trixter

Trixter — Mon, 17 Dec 2012 18:50:29 +0000

Of course!!! (smacks head repeatedly)

You have the new record! Congratulations sir!

By: Peter Ferrie

Peter Ferrie — Mon, 17 Dec 2012 18:44:29 +0000

12.
init: mov ax,0e5ch
scasb
pickch: jp writec
mov al,’/’
writec: int 10h
jmp init