linusakesson.net

Forum comments in chronological order

Disclaimer: I am not responsible for what people (other than myself) write in the forums. Please report any abuse, such as insults, slander, spam and illegal material, and I will take appropriate actions. Don't feed the trolls.

Jag tar inget ansvar för det som skrivs i forumet, förutom mina egna inlägg. Vänligen rapportera alla inlägg som bryter mot reglerna, så ska jag se vad jag kan göra. Som regelbrott räknas till exempel förolämpningar, förtal, spam och olagligt material. Mata inte trålarna.

Nov 2014

GCR decoding on the fly

Anonymous
Wed 5-Nov-2014 20:31
Awesome work.

In interest of pure speed, I would love to see this used with a non-IRQ/no-display loader for 2 revolution track transfer. I've looked into it and don't think it's possible on highest density tracks unfortunately.

From the end of this processing until the sector header 2 away you have:
- current sector tail - 2 00s + 4 or more gap = 6+ bytes
- skipped sector header - 5 sync + 10 contents + 9 gap = 24 bytes
- skipped sector data - 5 sync + 325 data + 4 or more gap = 334+ bytes
- for total of 364+ bytes (note these are all in terms of encoded (GCR) not decoded bytes)

So 364 is the minimum number of disk bytes that may fly by in the time you need to transfer the sector just read. After that you need to go back to reading a sector header, etc.

At highest density setting, this is 26us per byte, or 26 * 364 = 9464us.

The fastest serial bit-bang transfer I think has ever been done is Vorpal V1 (like used in Winter Games). It transfers 3 bytes in a 113us loop, for 37.67us / byte.

We need to transfer 257 bytes (256+checksum) in 9464us. This means each byte must be transferred in 9464us / 257 = 36.82us.

So I'd expect Vorpal style serial transfer to allow you to squeeze by with 2 revolutions per track on lower-densities, but just miss the mark and still require 3 revolutions per track on highest density.


I would still love to tweak this code to be more density dependent to eliminate that BVC or two though. That should free enough cycles to replace PHA with STA $1801 and allow a (non-IRQ/no-display) 1 revolution parallel port track transfer :).

Aaron

Autosokoban

Anonymous
Tue 11-Nov-2014 17:25
http://www.linusakesson.net/games/autosokoban/?v=1&seed=1093093105&level=12yy
Enjoy. (It's impossible.)
Not impossible. 21 moves. Nice puzzle

Larghetto from Chopin's 2nd Piano Concerto

Anonymous
Wed 12-Nov-2014 09:30
Is it your performance on the piano?

The TTY demystified

Anonymous
Fri 14-Nov-2014 01:31
Fantastic article. Thanks for doing all of the legwork. :)

-pf

Larghetto from Chopin's 2nd Piano Concerto

lft
Linus Åkesson
Fri 14-Nov-2014 12:00
Is it your performance on the piano?

Yes, it is.

The Chipophone

Anonymous
Fri 14-Nov-2014 21:15
Linus, please send me full documentation on my e-mail: alex.qrp.1997@gmail.com

The TTY demystified

Anonymous
Sun 16-Nov-2014 02:37
Thank you so for the article!
And now, I know the meaning of "tty" and "pty"!

Maya2003

Autosokoban

Anonymous
Fri 21-Nov-2014 13:00
Looks like it eventually runs out of levels at level 2147483647.
(I don't know much about the language you programmed it in, but is there a reason that the level counter is signed? However, that would just delay the problem...)

It's some sort of overflow I'd say as 2147483647 = 0b01111111111111111111111111111111 (so one zero and 31 one's)

(foto)

Anonymous
Tue 25-Nov-2014 15:57
👍

(foto)

Anonymous
Wed 26-Nov-2014 14:01
👍

(foto)

mporshnev
Max Porshnev
Wed 26-Nov-2014 14:39
55°42'12.08"N 13°11'57.88"E ?
lft
Linus Åkesson
Wed 26-Nov-2014 15:32

mporshnev wrote:

55°42'12.08"N 13°11'57.88"E ?

Correct!
mporshnev
Max Porshnev
Thu 27-Nov-2014 08:23
Tack, lft.

SID theme search engine

Anonymous
Sun 30-Nov-2014 03:03
Very nice little thing. Way better than other search engines in my opinion as you can control the input better. And the database is whopping large!