CPU/Processor Upgrade for the Atart ST

[alexh]

I think the Apollo team must have failed to find a good quality expert Atari user to help develop the Apollo Core using a Vampire A500 v2 on the Atari ST?

Instead they have created an MMU configuration which mimics the Atari ST memory layout allowing them to use EmuTOS and FreeMinT on an Amiga!!

http://www.apollo-core.com/knowledge.ph ... 9&z=04xteM

[Cyprian]

IIRC Vampire v2 doesn't support Atari ST due to lack of signals like VPA,VMA and E needed by ST

Below you can fin Majsta statement from 2013: "This accelerator can't work with Atari STE because he uses peripheral bus for older devices. So signals like VPA,VMA and E are used. Those signals are not used on A600. In A500 they are so it could be done for any computer who uses MC68K but since I don't have Atari and admire only Amiga computers I won't start projects related to Atari or any other manufacturer but I m ready to help anyone who want's to start similar project"

and

@bilt yes like I said code and everything is prepared to Vampire500. E,VMA,VPA generated any maybe in the near future I will show how they work. but like I said I don't have passion for any computer except Amiga. I would never create anything for any other computer. There is nothing interesting there. Only Amiga...

[AnthonyJ]

IIRC Vampire v2 doesn't support Atari ST due to lack of signals like VPA,VMA and E needed by ST

I guess there may be differences between the v2 and the v4 in this respect then. I'm fairly sure I can read the text "Vampire_4" on the board connected to Tuxie's ST in the photo just a few posts up ( http://atari-forum.com/viewtopic.php?f= ... 25#p342576 ), and it's quite clear that he/they are working on making it boot.

The "you can use v4 + 2.9 core in Atari mode in an Amiga" news is just another step forward along a related path, and is allowing implementation of the Atari video modes prior to having it actually booting in an Atari. The progress here briefly made 1st1 mistakenly believe that Tuxie's ST was booting which it isn't yet - but I imagine it will get there (and have similar CT63 beating performance as when running inside an Amiga host).

[Cyprian]

true.

anyway, I was waiting 5 years and actually I'm ready to wait next 5y for it

[joska]

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.

The fact that applications are crashing with memory protection enabled is a sure sign that memory protection has a huge benefit. Or a sign that your setup is faulty.

My fastest real machine is a Firebee. When you're developing on the actual hardware it's a huge pain to debug without memory protection, so I stopped using it a few years ago. Went back to the "slow" Milan060, which despite it's "slowness" offers a much more productive environment. So clearly, until the Vampire gets a real MMU it is of no interest to me.

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.

This will work fine on a Falcon, not so fine on an ST. If I understand it correctly you will be able to switch back to the original CPU when needed, in this case you must either fall back to the original shifter or implement the shifter in every tiny detail on the Vampire in order to run games and demos that use shifter tricks.

On the ST it would be better to route RRRGGGBBB from the shifter to the Vampire in order to display actual shifter output. The Vampire "shifter" would only have to support shifter bitplane layout to allow the TOS internal VDI (or NVDI) to use larger resolutions set up by a separate driver (or patched TOS/EmuTOS).

Still tomatoes on your eyes...

Your tomato- and spaghetti-jokes may be funny in your native language, but in English they just sound stupid and quite offensive. Your posts displays your limited experience with and/or understanding of the TOS platforms, yet you like to talk down anyone who criticise or correct you.

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.

A very good example of your offensive and obnoxious posts. You are probably the only one who has done damage to the Vampire ST project, by offending a lot of people that could have contributed.

[1st1]

Next step seems to have been done, SAGA can support ST Shifter video modes. http://www.apollo-core.com/knowledge.php It's not clear from that link if that is running already on a ST mainboard or Amiga.

[alexh]

IIRC Vampire v2 doesn't support Atari ST due to lack of signals like VPA,VMA and E needed by ST

That is not correct.

Below you can fin Majsta statement from 2013: "This accelerator can't work with Atari STE because he uses peripheral bus for older devices. So signals like VPA,VMA and E are used. Those signals are not used on A600. In A500 they are so it could be done for any computer who uses MC68K but since I don't have Atari and admire only Amiga computers I won't start projects related to Atari or any other manufacturer but I m ready to help anyone who want's to start similar project"

That pre-dates Apollo core. That is for the original Vampire FPGA for A600 only. The Vampire A500 v2 seems to have all the signals required going between the socket and the FPGA via level shifters.

[alexh]

Next step seems to have been done, SAGA can support ST Shifter video modes. http://www.apollo-core.com/knowledge.php It's not clear from that link if that is running already on a ST mainboard or Amiga.

It's on an Amiga.

[1st1]

IIRC Vampire v2 doesn't support Atari ST due to lack of signals like VPA,VMA and E needed by ST

That is not correct.

Now you are not correct. In the first tests a few weeks ago in ST board they had to rewire some Vampire test pins to unwired pins on the 68000 socket. We told the Apollo team already a year ago or so that they should wire these, but they didn't. The processor core itself internally supports these signals since long long time, but sourrounding SAGA infrastructure did not linked them to the socket. So they redefined some test signals and changed their logics and linked them to the missing signals on the socket for testing ST. My last information was that there is currently some MFP interrupt handling which is handled incorrect that ST can boot. I have to check for newer status.

[alexh]

IIRC Vampire v2 doesn't support Atari ST due to lack of signals like VPA,VMA and E needed by ST

That is not correct.

Now you are not correct.

Are you sure? 4 days ago Igor Majstorovic (creator of the Vampire boards) confirmed these signals are on the Vampire 500 board.

http://www.apollo-core.com/knowledge.php?b=5&note=14252

In the first tests a few weeks ago in ST board they had to rewire some Vampire test pins to unwired pins on the 68000 socket. So they redefined some test signals and changed their logics and linked them to the missing signals on the socket for testing ST.

That sounds like something you might have to do with a Vampire 600 board but not a 500 board.

[1st1]

So how you would plug a Vampire 600 into a Mega ST? Ingo and the Apollo Team do their tests with ST and Mega ST, not STE.

[joska]

Are you sure? 4 days ago Igor Majstorovic (creator of the Vampire boards) confirmed these signals are on the Vampire 500 board.

http://www.apollo-core.com/knowledge.php?b=5&note=14252

From that post:

V500 don't have bus arbitration logic signals(BR,BG and BGACK)

So the V500 will not work in an ST.

[alexh]

V500 don't have bus arbitration logic signals(BR,BG and BGACK)

So the V500 will not work in an ST.

Interesting stuff. Unfortunately I don't know enough about Atari ST/ 68000 bus architecture to know if those are required in an ST where the 68000 has been removed.

The reason they are present on Vampire 600 and not on Vampire 500 is because on the Amiga 600 the CPU cannot be removed and they are used to take over from the 68000 on the motherboard. But the Vampire A500 didn't need them because the 68000 is removed.

In the first tests a few weeks ago in ST board they had to rewire some Vampire test pins to unwired pins on the 68000 socket.

I wonder if BR, BG and BGACK are the signals you were referring to?

[joska]

Interesting stuff. Unfortunately I don't know enough about Atari ST/ 68000 bus architecture to know if those are required in an ST where the 68000 has been removed.

I'm quite sure they are. They are used by the GLUE to allow the CPU access to the bus.

Another thing - does the Vampire support the processor status pins? These are IIRC not used in the Amiga, but they are in the ST.

[alexh]

I'm quite sure they are. They are used by the GLUE to allow the CPU access to the bus.

Interesting.

Another thing - does the Vampire support the processor status pins? These are IIRC not used in the Amiga, but they are in the ST.

FC0-2 & IPL0-IPL2 are raised on the Apollo forum as "Unused by Amiga" in addition to the pins already discussed. There is talk that the Vampire 600 V1 did not have them but nothing about Vampire 600 V2 or Vampire 500 v2(+)

As three pins (BR,BG and BGACK) are already missing from the Apollo 500 V2(+) if there are any more there are unlikely to be any spare level shifter pins.

[1st1]

That's what I wrote, the testing in ST currently works with hand modified Vampire cards. Vampire board has some test pins. These were redefined in the FPGA programming to behave like the missing signals and connected to the unused 68000 socket pins. As soon as it works, I think they will modify the board layout to link these former test signals directly to the socket. V4 board is still not available, maybe this is the reason, or at least one of the reasons.