Welcome, Guest. Please login or register.

Author Topic: A new Napalm speedup hack, help needed  (Read 5345 times)

Description:

0 Members and 1 Guest are viewing this topic.

Offline Skeleton

  • Newbie
  • *
  • Join Date: May 2005
  • Posts: 1
    • Show all replies
    • http://www.metasyntactic.se
Re: A new Napalm speedup hack, help needed
« on: May 07, 2005, 10:32:02 PM »
Quote
Can anyone get in touch with Kalms or Marusa


Sorry for the delay folks. Registering on the forum gave me a semi-hard time.
Anyways, I have been in touch with Mr Kalms and this is his reply:

Quote
I disassembled and checked the Suzy c2p out.


It contains a somewhat older version of the C2P routine inside
c2p1x1_8_c5_030.s (the latest version I published has a few instructions
shuffled around), which has been modified to support arbitrary
bitplane-sizes, and then the RTG C2P header has been tucked onto its
front.


If you are targetting 040+, you probably want to use another routine,
because the Suzy routine reads the source data twice during C2P
conversion.
(It is optimized for 68020/68030.)


All my routines are in the public domain (modify them, sell them for
profit - I don't care). If you want to be absolutely 100% in the clear,
it would be quicker & easier for you to take some other C2P, implement the
deltaskip optimizations there, and make your own RTG C2P package out of
it.
I would suggest that you try with c2p1x1_8_c5_040.s, because it already
supports arbitrary bitplane sizes.


You can find c2p1x1_8_c5_030.s and c2p1x1_8_c5_040.s inside my C2P archive:


http://www.lysator.liu.se/~mikaelk/src/kalmsc2p.lha


-Mikael Kalms