Re: [MiNT] Shutdown() discussion

["Konrad Kokoszkiewicz" <draco@obta.uw.edu.pl>]

>> You need speed? How move.w #xxxx,$yyyy.w is slower than an illegal
>> instruction or a JSR?
>
> JIT compiler... Look at speed of IDE and hostfs under JIT compiler.
> You see how it is slower.

Obviously, I keep forgetting about JIT.

> OKi, dou you know some possibility? Send it to ARAnyM's developer
> list, please. Or we can look at the thread you mean.

Xavier's idea is worth respecting IMHO. No cookie jar entries, pure illegal
opcodes. Aranym must make them to be privileged (so that only the OS/drivers
could use them) and that's all. /dev/aranym is the answer when API is
concerned.

Of course, you can answer "TOS, MagiC", but a version of /dev/aranym can be
developed for Magic (it has an idea about devices), and about TOS support
let us let care them, who use it. At last, TOS cannot offer all the
possibilities MiNT offers. As Frank said, if someone needs something MiNT
offers, always can boot MiNT. This is MiNT list, TOS is out of our scope,
TOS issues cannot force MiNT to be inconsistent. Moreover, we are sorta
interested to convince remaining TOS users to switch to MiNT (which is now
TOS, but the way ;)).

> Really is it so slower? On my machine not. X-mas comming, so the time
> to look at ARAnyM for me.

The MMU version is seven times slower on my machine without JIT, than
non-MMU version without JIT, to be strict. Of course, it may depend on the
host CPU speed and motherboard speed.

--
CVV
Konrad M.Kokoszkiewicz, http://draco.atari.org

** Ea natura multitudinis est, aut seruit humiliter, aut superbe dominatur.
** Taka to juz natura pospólstwa, albo sluzalczo sie plaszczy,
** albo bezczelnie sie panoszy. (T. Liuius XXIV, 25).