Re: neue Email

From: Johan Klockars <rand_at_cd.chalmers.se>
Date: Sat, 25 Jul 1998 09:06:48 +0200 (MET DST)

OK, I have now checked out my slightly hacked Frontier.

It appears I was somewhat wrong. It is indeed the instruction cache that
causes problems, but see below.

> >Does anyone know how to fix the game 'Frontier Elite 2' to run it with both
> >caches ON with the Afterburner040? It seems the game doesn't like the
...
> Erm, no. I think Frontier is one of those advocates of self-modifying code,

I don't think so (unless you count the depacker).

> the result being - complete failure on a 68040. :)

But I don't get a complete failure here!

Apart from two things, the game works well in FastRAM (I simply set the
proper flags on the unpacked executable) since it uses fixed addresses
for its screens (f0000 and f8000, which may well overwrite your disk
caches and such things (I wrote a small AUTO program that allocates a
large block of memory to circumvent this)) with both caches on:
- The mouse movement problem I described, where movement is often
  interpreted as key presses instead.
- The palette changes aren't done at the correct time, causing strange
  looking flicker in the image.

There is a very noticable difference in the update frequency between running
the game with both caches on versus with just the data cache.

Anyway, I'm pretty sure this is the MOVEM problem or a variation on it.
There is indeed at least one large block of MOVEMs (16 or so with a whole
bunch of registers) in the GAME and we now that can cause interrupt problems.

-- 
  Chalmers University   | Why are these |  e-mail:   rand_at_cd.chalmers.se
     of Technology      |  .signatures  |            johan_at_rand.thn.htu.se
                        | so hard to do |  WWW/ftp:  rand.thn.htu.se
   Gothenburg, Sweden   |     well?     |            (MGIFv5, QLem, BAD MOOD)
Received on lø. juli 25 1998 - 13:40:00 CEST

This archive was generated by hypermail 2.3.0 : ti. nov. 03 2015 - 20:07:54 CET