To: vxp@imaginet.fr ("centek" )
From: dlanor@oden.se (Ronald Andersson)
Organization: Atari
Subject: Re: Full turbo speed for MagiC (solved)
X-Mailer: NEWSie Version 0.94 (Atari)

>Hello,
>
>MAGIC runs in FAST-RAM with TOS 7 !!!!!!!!
>
>We have 170 CT2 running and I think that about 50 % use magic (in tos 7,
>sure !!)

That sounds good, but I still have to do it with my own CT2B to feel
perfectly satisfied.


>But what you don't understand is that your CT2 has a problem with
>timings in Fast RAM !

Why would I not understand that now, all of a sudden...?

I was the one who first suggested it, in some of the earlier mails
of the many I have lately exchanged with Centek staff.  I just don't
agree that this particular problem could cause my MagiC symptoms.

Such timing problems should affect the FastRAM access the same way,
regardless of when the system begins to use the FastRAM.  That is where
my experimental results are interesting, because I did succeed in using
MagiC completely reliably with FastRAM after booting in TOS 4.x mode,
even though I could never get such reliability (or even get near it)
when booting in TOS 7.x mode.

The timing problems probably explain the overheating of my CT2B, and so
the other problems that depend on heat.  But they do not explain the
great improvement in FastRAM reliability when activated in TOS 4.x
mode by my patch, as compared to the bad reliability under TOS 7.x.


>So, all your test with magic have NO value !

I disagree. They proved beyond any doubt, that the very same FastRAM
being addressed by the very same CT2B worked excellently when MagiC
was booted in TOS 4.x mode and FastRAM was activated by my patch.

That does indicate a rather important difference between TOS 7.x
operation and TOS 4.x operation.  The cause of that difference may
lie either in hardware or software design, we can't know that yet,
but we _do_ know that a difference exists.  That is not 'NO value'.


>You lost your time with patch ....
>
Not really, even if you were correct in all of the above.

I got a couple of days worth of using my machine with MagiC at full
speed that I could not get in any other way at the time.  That alone
was worth the fairly short time it took me to make the simple patch
program. After I got the idea for it, it only took a few minutes to
code it, so I actually think I gained some time rather than losing any.

The real time loss is in waiting for 'snail-mail' delivery...
(Like now: I can't get the CT2B until tuesday next week, at best.)
