One
Year Of CT60
by Anders Eriksson
At the time of
writing, the CT60 is now shipping in its second
batch. It's time to look back and see what all
the fuss has been about, why our precious forums,
on-line mags and news sites a have been filled
with articles about this little add-on.
One year ago It
has been a year since the first 30 (approximately)
beta developer cards were shipped. The problems
emerged instantly. The beta cards had quite
a lot of firmware bugs. Unfortunately Czuba-Tech
wasn't very clear about the shipment, so the
majority of the developers weren't aware about
them being a beta-tester. Quite some complaints
all over the place, which must have been rather
frustating for Czuba as very few seemed to realize
this wasn't the final product.
To clear up some
mess, and to centralize the discussions we opened
the CT60 part of dhs.nu, including a discussion
forum. This part of the site is (still) maintained
by Deez of Evolution.
New firmware
upgrades arrived steadily, with one bug after
the other ironed out. Luckily Czuba-Tech had
chosen a programmable logic so the upgrade of
the accelerator could be done via software,
a very elegant and painless solution. At the
point where Czuba couldn't find any more bugs
in the firmware, things were still working really
bad at some places (including mine). Programs
with a size larger than 70 KB crashed during
execution! This was a real mind-bender and everything
was weird to say the least. In the end I downgraded
HD Driver from 8.0 back to 7.x, and voila, all
problems were gone.
Those 30 seconds
of a job to install the older HD Driver had
caused weeks of debugging and head-scratching.
Ah well, I suppose that's life when you're beta
testing new hardware. Now, the known problems
were all in the past, and we're around late
July 2003. Czuba-Tech started to ship the boards
to everyone.
Shipment in
progress The cards were shipping at long
last, and just a couple of weeks after the deliveries
had begun, a new firmware bug was detected.
Now, this wasn't very good. Non-beta cards had
been sent, and they needed firmware re-programming.
I think everything went fairly smoothly thanks
to the beta testers who already had the knownledge
and tools (JTAG) to re-program the chips, luckily
only a handful of cards had left with this bug
intact.
Things calmed
down and the cards were delivered one after
the other. We heard stories about installations
from all over the world, with surprisingly few
issues reported (compared to older boards such
as the CT2).
New 68060
chips! During late fall 2003, Czuba-Tech
surprised everyone by reporting a new 68060
chip which was much more tolerant to high frequencies
than the old revisions. All developers were
offered the chance to replace their old CPU
chip with this new one. We started to see 100
MHz and faster Falcons just a few weeks after
the announcement. Some maniacs went above 100
MHz with water-cooling systems!
2004 - more
CT60s? Before even all of the CT60s had
been shipped, talks about a second batch were
happening. In February the next batch was announced,
80 new boards were to be built, giving a total
of 230 CT60s out there. And that's where we
are today, the first shipments of the new CT60
batch has arrived and we could possibly say
that the CT60 has been stabilized in the Atari
community now.
CT60 specific
programs? These are rare, except the
CT60 configuration tools that is. To date we
haven't seen any applications specifically for
the CT60, only a few patched programs. Now why
is this? First of all, and the most important
reason is that 68060 code is backwards compatible.
Which means a CT60 game would also run on a
68030 or 68040 Falcon, but possibly much worse.
One of the never-ending
benchmarks is Quake. Sure it will launch on
a bare Falcon, but you can count to ten seconds
between each screen update. It's clear that
one is not designed for a bare machine.
Still, these
programs do not count as CT60-specific programs,
even though they are practically only usable
with CT60. Other programs such as pixel-painters,
soundtrackers, text editors, assemblers, word
processors and so on work as-is on the CT60,
and these kind of applications need no special
68060 fancy routines. They just work faster.
Much faster.
So, how about
demos? Demos always find a way to use the new
hardware in some way the normal applications
just never succeeds. To date there are zero,
nada, no demos at all made for the CT60 (not
counting a few smaller intros). Why, oh why
is that? One of the major reasons is of
course that only a few of all those who bought
the CT60 seems interested to make a demo for
it at all. Long-time Falcon coders who bought
CT60 have been dissing it becuase the only "true"
coding is on a bare Falcon. This leaves only
a few behind who don't care for these ideologies
and just wanna code for fun.
Another reason
is that the entry level quality of a CT60 demo
is expected to beat everything previously made
on the Falcon, preferably by miles. Both in
design and technical aspect. We have all seen
the latest and greatest Amiga 68060 demos, right?
Tell you what, this is no easy thing to have
on your throat when you're coding on new hardware
for the first time. The Amiga fellows have had
nearly a decade to learn the 68060 chip and
how to tweak the most performance out of it.
Needless to say, the Amiga coders have also
had the chance to do the "modern"
style demos for a much longer time with complex
3D scenes and the likes. It will take a long
time for us to play catch up with that.
If the bar was
lower, I'm sure we could have seen some simpler
demos out already, like we did for the original
Falcon. Many of the early ones were ST code
beefed up a bit. But somehow the expectations
seem sky-high and that's an important reason
why (the few) CT60 demos are yet to be seen.
The future The
time will come, when the CT60 demos appear for
real, that's a promise. If you visited the Outline
2004 party, you could see an early preview of
the upcoming Evolution demo for the CT60. Take
a seat, and hold on to your hats, there will
be action.
Some other intersting
ideas for the future come from the land of Amiga.
Namely, porting Amiga demos to CT60. Ephidrena
seems interested to try this, and TBL has been
investigating as well (including trying out
Atari compilers). If the Ephidrena and TBL productions
make it, we have some near-impossible benchmarks
to match. TBL and EPH are years ahead of anything
we've yet seen on the Falcon. With extremly
talented 3D graphic artists (something the Falcon
scene is completely missing out on). Their 3D
scenes look very impressive even if the underlying
code is on the same level as what we (will)
have on the CT60. The art of modelling in low-polygon
detail is a new world for many Atari artists
to try. A few toruses, spike balls and cubes
just don't cut it any longer.
So in short,
the future looks quite bright, possibly with
some very impressive things happening, either
from the Atari coders or from the Amiga ones.
Will the only breathing 68K platforms become
more united than ever?
Why not, as Synergy
said back in '93: Diversity brings chaos.
|