CPU/Processor Upgrade for the Atart ST

[Foxie]

TF530 - will not work on STE... (TF530 dont like blitter)

I suppose that means it won't work on the Mega either? It would be much easier to max out my Mega than my 1040STFM - it already has 4MB and can easily take a graphics card. I'm not even sure how feasible it would be to fit a graphics card to a 1040. Lots of soldering.

What I'd really like is a switch to revert back to the 68000. Do you think that could be done by putting either the accelerator or the 68000 into reset depending on the switch position?

Perhaps the blitter could be somehow disabled at the hardware level when the accelerator is switched on. TOS must have some method of detecting the presence of the blitter.

[joska]

TF530 - will not work on STE... (TF530 dont like blitter)

Why doesn't it like the blitter?

[DarkLord]

Yes, with patched games there is little incompatibility. But forget anything with syncscrolling or border removal unless you have a real 8MHz 68000 available.

I think you're right. Just for the fun of it though (and because I'm a glutton for punishment ), throw out a
handful of titles and let me try them so I can watch them crash and burn. Thanks.

[Foxie]

I think you're right. Just for the fun of it though (and because I'm a glutton for punishment ), throw out a
handful of titles and let me try them so I can watch them crash and burn. Thanks.

Enchanted Land would become a mess. Obsession might be interesting. It does a slight horizontal overscan using an STE shifter trick. I'm not sure if that trick needs sync code. Of course demos will mess up terribly. I tried running Sea of Colour under an emulated MegaSTE, and you can see the effect of the extra blitter waitstates.

Now the real challenge would be to patch those games and demos to work on an 030, 040, and 060 at common clock speeds. If you turn the cache off, you should be able to predict the timings. But I'm not familiar with 030+ timings. On the higher CPUs things like branch prediction could turn that into a complete nightmare!

[DarkLord]

Well, the STacy won't run any STe stuff....

But I'll try Enchanted Land (assuming I can find a patched version here)
and see how badly it goes later on tonight.

Thanks.

[ranix]

I don't think a CPU upgrade necessarily needs to be able to run with the base specifications. Sometimes a performance increase is worth setting up another console for

[1st1]

There are other accelerators with faster CPU's (020 and 030), I have no real experience with them. However, it's clear that compatibility suffers compared to a 68000 accelerator, unless there's a possibility to fall back to the original CPU. No problem if you don't play games though. If you're a "serious" user then these types of accelerators are probably the best choice. The Vampire - if it ever arrives for an Atari - would fall into this category.

Again!!! We are working on it !!!!

Tuxie, the international Atari users have tomatoes on theyr eyes, they have beans in their ears, but they have big mouth in not believing that it happens. Unfortunatelly, since the announcement of Apollo team to do it and requesting help to assist it to get it working, here in the forum they (and myself) have been mocked. We lost a year. We lost a year. We lost a year.

Stay tuned.

http://www.apollo-core.com/knowledge.php?b=6&note=13491

apollonews.jpg

I told you so. We could have had it already since a year ago. We lost a year. We lost a year because of skeptic mocking by you guys.

Thanks to Tuxie and Gaga to assist them (Apollo team) with their knowhow! You are the heroes. Soon we will see it alive.

[damanloox]

Is anyone working on something like this for the Atari Systems?

Natural beauty doesn't need any makeup and Atari love their machines as they are )

But seriously - CT60 implemented in some sort of FPGA would be great addition to falcon (or - even better - the whole ST/TT/Falcon family). Looking at all the Vampires implementing 68060 in FPGA seems to be possible...

[joska]

The Vampire - if it ever arrives for an Atari - would fall into this category.

Tuxie, the international Atari users have tomatoes on theyr eyes, they have beans in their ears, but they have big mouth in not believing that it happens.

And some people clearly have tomatoes and beans for brains.

Good news, but until a real MMU appears it is still of no interest to me.

I told you so. We could have had it already since a year ago. We lost a year. We lost a year because of skeptic mocking by you guys.

Brave words for someone who has done nothing than blabbering about stuff you have no knowledge about.

[ggn]

told you so. We could have had it already since a year ago. We lost a year. We lost a year because of skeptic mocking by you guys.

Nope, only you lost a year. We only spent one year LOLing about how obnoxious and clueless you are .

[1st1]

Good news, but until a real MMU appears it is still of no interest to me.

Tryed MMU support on TT & Falcon-CT60 with EasyMint. Result: More crashs with applications. It's nice to have, but benefit in MiNT is small. Anyhow, as soon as Vampire runs in ST, the more there will be requests for MMU. Step by step.

[tuxie]

We lost a year. We lost a year. We lost a year.

Stay tuned.

http://www.apollo-core.com/knowledge.php?b=6&note=13491

apollonews.jpg

I told you so. We could have had it already since a year ago. We lost a year. We lost a year because of skeptic mocking by you guys.

Thanks to Tuxie and Gaga to assist them (Apollo team) with their knowhow! You are the heroes. Soon we will see it alive.

This is not true, the core was not done that it would ne able to run on ST. The maprom function is done since only few monthes, and its still hard to bring the Vampire up on ST.
What is done already
EmuTOS implemented into Maprom and mapped to E00000
Berr implementation is done but needs some optimizing
Reset adaption is in work (Amiga has no internal CPU reset, so every Tos Reset call the cpu is resetted to so a loop occurs)
IRQ Autovector adaption - wip
FC0-2 adaption - no started
Bus request - no started
so there is still alot of work.

What we did is to optimize the Saga driver.. our goal is to beat the SuperVidel benchmarks

[1st1]

Still tomatoes on your eyes...

http://www.apollo-core.com/knowledge.ph ... 1&z=53Dmhg

Gunnar von Boehn
(Apollo Team Member)
Posts 3253
31 Mar 2018 11:59

1ST1 wrote:
But this is still an A500 board as the base... In the Atari scene there will not be much acceptance until it is alive with an ST board as the base.

Yes we work on different things:

1st)
All AMIGA+Vampires can with upcoming GOLD 2.9 with 1 click boot into EMUTOS / MINT.
This means AMIGA+Vamp can be compared very much to a Firebee in regards running ATARI OS.

2nd)
Yes you can plug a VAMP physically into an ATARI ST.
The smoke test we survived already.

We are testing software, meaning the booting of TOS and EMUTOS right now. The booting of the Vamp in real ATARI is at an early state, we gets Bombs still.

3rd)
We started speed tuning the ATARI SAGA VDI driver.
Some functions already run 50% faster than before.
Here is great potential for more tuning,
and VAMP can soon be one of the fastest GFX solutions for ATARI.

I think in the near future VAMP can be used by ATARI owners for maximum speedup.

Credit for our progress goes to Ingo, who is helping us understand ATARI OS better.

Cook some spaghetti and use the tomatoes for sauce.

This year we will see real ST hardware running with Vampire. And it will outperform anything elese what can be plugged into ST and even ob Falcon board. We already lost more than one year, also because of skepticism in this forum.

Remark: Bombs is already something, that means, TOS passed some early tests like memory and accessing and initializing the Shifter. They are on a good way.

[troed]

Sometimes you might want to check who you're replying to _before_ you press send.

[tuxie]

I have attached a shot of my ST
https://1drv.ms/u/s!AuWW5C3ZWdkMi3GXB4d9xxOOCew2
https://1drv.ms/u/s!AuWW5C3ZWdkMi3C9RaamG_8VrsU-
Tos3.06 (PAK TOS)

[1st1]

Sometimes you might want to check who you're replying to _before_ you press send.

I didn't reply on Tuxie. See it as a news update.

[1024MAK]

Play nice please. This may be the chat area, but there is no need to be disrespectful to other members. If you don't have anything useful to say, or your criticism is not a constructive comment, then refrain from posting.

If members can't behave, then the moderators will take action.

Mark

[wongck]

I don't see why one will not do it for the Atari.... it's an untapped market for them.
Otherwise, why they come over and ask for help one or more years ago?
So long as there are people working on it, like anything else, they will overcome the obstacles and make it happen.

[1st1]

The SAGA continues... Apollo teams is asking for informations about video modes of ATARI chipsets. Do you remember, a year ago I already asked in this forum for exactly this kind of support to them from specialists logged in here.

They sooon will have booting ATARI ST with plugged in Vampire V4 card, and next step will be to implement the emulation of ST/STE/TT-Shifter and Videl into SAGA graphics core. Like SAGA also emulates Amiga OCS/ECS/AGA graphics modes as well, this would be specially important for the stand alone version, but also for accelerator card, as you don't need to change the monitor to use original resolutions and SAGA in higher resolutions, it would all come out of one video connector. I think they would need support for non-documented tricks like overscan support to get really compatible, the basics they can read from Profibuch or similar sources. I sugested them to study Hatari emulator and MiST Atari ST core as a first step. But anyhow, maybe they need some assistance to fully understand what's going on there. On the other hand, as such tricks like Overscan are very timimg sensible, and a CPU turbo card never is cycle exact, existing software to use these tricks anyhow would not run properly. Difficult decision if it's worth to implement the design flaws of ST/STE shifter...

See http://www.apollo-core.com/knowledge.ph ... 9&z=Uz3V-K

[troed]

No, there's no need to emulate 68000/GLUE cycle exact tricks on an accelerator.

[Cyprian]

1st1 is wrong

Just to clarification, let me cite Gunnar von Boehn:
"The question was which GFX formats to real ATARI chips support."
More you can find there: http://www.apollo-core.com/knowledge.php?b=5&note=13959

[1st1]

Where I wrote something different? I just make them aware of undocumented resolution tricks and open discussion if it it worth to implement them or not. Finally you are linking to the same thread as I did.And I argumented there exactly the same way.

[1st1]

Looks like, they got it... The baby seems to fly... Falcon with ct60 is a slow (not lame!) duck against this... (except some graphics benches on SuperVidel, but this seems to be possible to speed up by optimze the SAGA-VDI driver)

See http://www.apollo-core.com/knowledge.ph ... 4&z=zNrkDo

But it's not ready yet, Apollo team next will implement ATARI Shifter (ST/STE/TT) and Videl graphics modes into SAGA. I think STE/TT and Falcon compatible audio hardware also should be added (Vampire has audio hardware, and I think they already implemented Paula).

[Cyprian]

1st1 take a deep breath and sit for a minute.

There is no Vampire card for Atari yet, therefore nothing has changed for us.

I track that project since 2013 and I can tell you that Gunnar von Boehn Majsta do a great job. Unfortunately still not for us.

[1st1]

Believe me, they are working on it. Above you see a posting from Ingo, they are doing it.

But it's still too early, I hoped so, but it is still not booting successfully. But we soon will see it.