Jump to content
IGNORED

Apollo "68080" new 68K Core running on Atari STf


Xebec

Recommended Posts

TL;DR An accelerator with the performance of a 68060 @ 150 mhz running in an Atari STf.

 

....

 

I'm not sure how familiar folks are with the "Apollo" CPU core and the "Vampire" line of accelerators for the Amiga. The Apollo CPU is a ground up brand new 68K compatible core running in FPGA, current performance in a relatively low price FPGA (large but not super fast) makes the CPU core at least as fast as a 68060 @ 150 mhz (fastest speed that shipped was 75 mhz). The Vampire accelerator uses the "Apollo 68080" core and via the FPGA also adds AGA chipset (and more features) to older Amigas such as the 500, 600, etc. The apollo core also has instruction compatibility across all 68K cpus so should be more compatible with the original 68000 than the 68040 or 68060..

 

Here's a post that shows the Apollo team has a Vampire accelerator running in a 1040STf:

http://www.apollo-core.com/knowledge.php?b=1&note=1962&order=&x=1

 

Scroll down to about the bottom and you'll see the pic.

 

It looks a little funny because they stacked a bunch of 68000 CPU sockets in there to raise the board above the shielded part of the ST chipset (the video shifter was in the middle underneath the shielding?). This is pretty exciting as this would literally take a stock STf and give it more CPU performance than a 68060 accelerator would. (The Vampire accelerator also comes with 128MB of RAM onboard , "fast ram" for the FPGA CPU).

 

In the future they're going to do an even higher end version that raises the performance about 3x, but of course cost 3x. The current Vampire 600 accelerator goes for 250 Euros, and a faster binned version is 289 EU. They are ramping up production of the 600 version and the Vampire Amiga 500 version (shown attached to an Atari 1040STf in the pic).

 

P.S. To be clear i'm not associated with this group in any way, just super excited at their work :).

  • Like 2
Link to comment
Share on other sites

That's competently pointless with no use, no offense, i see your excited, but you should look at the new amiga vampire 500 boards that they are selling on ebay right now. They are going for $650-840 USD. I am very curious (and very doubtful) if there would be any atari st owners will buy one of those for that kind of prices? Is there any practical use for running an st @150mhz i might not be aware of? If an Atari fan wants to drop cash like that, then the Falcon is the goto system imo.

Edited by cjameslv
Link to comment
Share on other sites

I'm all for faster hardware on STE's or Megas or Falcons... sign me up. But here's a question to everyone. We know we can "go faster" with emulators, coldfire firebees, and various accelerators....but when does it become practical when new software to take advantage of these improvements seems limited?

 

Basically I think folks want to see their favorite games run faster. Some might want more power to do something with sound/midi processing and then their are those who like to tinker with old productivity software like DTP and graphics. The faster our machines get, the more likely these pieces of software will struggle to stay current. Heck... I'd love to have a 3D animation and rendering application on the Atari to create animations that is more capable and modern... but no one is going to write it. Are they?

 

What is the state of "new" Atari software outside of games or maybe music... anything? And does that require a 68080 card?

Link to comment
Share on other sites

That's competently pointless with no use, no offense, i see your excited, but you should look at the new amiga vampire 500 boards that they are selling on ebay right now. They are going for $650-840 USD. I am very curious (and very doubtful) if there would be any atari st owners will buy one of those for that kind of prices? Is there any practical use for running an st @150mhz i might not be aware of? If an Atari fan wants to drop cash like that, then the Falcon is the goto system imo.

 

These are fair points. Though i'll say the $650-840 is the scalping I mentioned in my post. They had very limited supply and some people were buying to resell. They've recently raised the base price but switched to automated manufacturing so the supply side will ease and settle at 250-289EU depending on the model (www.kipper2k.com).

 

So for the Amiga, ~$300 (250EU) gives you:

- A much faster computer

- Allows an Amiga 500/600 to run 1200/4000 AGA chipset software (maybe an ST version allows STf to become a STe?)

- SD card storage as a hard drive (replacing Ultrasatan possibly in an ST application, worth ~ $110)

- HDMI output (normally a $80 upgrade on the Amiga, not sure about ST)

 

There are people out there with coldfire FireBees - this might be a way to give them the ability to run TOS apps faster without spending the $600 on those. Depending on how the FPGA development goes, if this Vampire can allow an OCS/ECS amiga to run AGA software, it's not too far fetched to see a STf running STE software.. completely. The Amiga version also goes beyond AGA resolutions using a retargetable graphics driver, so the same might be possible on ST.

 

TOS Web applications that previously really wanted the horsepower of a TT/Falcon/Medusa/etc would also run really well on this (assuming they work on TOS 2.xx). Oh, and it's fun to watch an Amiga 600 now play MP3..

 

P.S. www.apollo-accelerators.com has some more details on the Amiga versions.

Edited by Xebec
Link to comment
Share on other sites

There is a bit of discussion about Vampire going on in one of the ST facebook groups I frequent, and the whole booster subject is of interest to me as I have one of these fitted: http://www.exxoshost.co.uk/atari/last/V1STE/index.htm

 

It's still a fairly standard 68K, but in all of my discussions with the creator of this booster, although the ROM and CPU will run at 32Mhz, the bus still runs at 8Mhz, so whilst I can understand the performance boost of the Vampire, I would question how it would improve access to existing architecture that's in place, unless of course it's adding those additional ports, and all the ST is doing is providing power and keyboard/mouse etc.

 

For 3d based games, I'm seeing about a 25% boost in framerate. I haven't tested much in the way of 2d games.

 

In terms of GEM, I'm seeing at least a 200% boost, and in some areas a 300% plus boost.

Link to comment
Share on other sites

 

These are fair points. Though i'll say the $650-840 is the scalping I mentioned in my post. They had very limited supply and some people were buying to resell. They've recently raised the base price but switched to automated manufacturing so the supply side will ease and settle at 250-289EU depending on the model (www.kipper2k.com).

 

So for the Amiga, ~$300 (250EU) gives you:

- A much faster computer

- Allows an Amiga 500/600 to run 1200/4000 AGA chipset software (maybe an ST version allows STf to become a STe?)

- SD card storage as a hard drive (replacing Ultrasatan possibly in an ST application, worth ~ $110)

- HDMI output (normally a $80 upgrade on the Amiga, not sure about ST)

 

 

 

STe graphics are not that much better than ST graphics (save for scrolling, but at that CPU speed, the CPU alone will be capable to improve scrolling if the game supports it). The de facto standard graphics expansion for semi-professionally used STs would be an ET4000 card (supported e.g. by NVDI 5). But where's the point? Hardly anyone will use Arabesque Professional, Chagall, Calamus SL, etc. any more. OK, but it will run DOOM. As does my N-Gage.

  • Like 1
Link to comment
Share on other sites

 

STe graphics are not that much better than ST graphics (save for scrolling, but at that CPU speed, the CPU alone will be capable to improve scrolling if the game supports it). The de facto standard graphics expansion for semi-professionally used STs would be an ET4000 card (supported e.g. by NVDI 5). But where's the point? Hardly anyone will use Arabesque Professional, Chagall, Calamus SL, etc. any more. OK, but it will run DOOM. As does my N-Gage.

 

The hardware is already developed for the Amiga; it sounds like it's pretty trivial to re-use the core on the ST. The FPGA code is already there for STe style hardware (FPGA-Arcade, MIST), so we're not talking a lot of work.

 

As to the point - do you use any of your old computers at all? Certainly emulation lets you run the same software and get some of the experience, but for some of us, there's nothing like using the real hardware itself, and imagining what could have been.. Maybe this isn't for you, but I eventually look forward to buying one of these.

Link to comment
Share on other sites

 

for some of us, there's nothing like using the real hardware itself,

 

The real hardware itself would be a 1040STE or Mega STE, with 8 or 16 MHz of CPU speed. These go for far less than this accelerator board, and I own about four or five of the former currently. As for usage, I do not use them for anything else than gaming, since the productivity stuff is done on this aging (Phenom II X4 910e) computer.

Link to comment
Share on other sites

Oh crap, this isn't going to devolve into one of those, "If it ain't original Atari hardware, it's

not an Atari", discussions, is it?

 

I certainly hope not. That tired old argument that if you do anything to your Atari its not an

Atari any more is pathetic and lame.... I'll never forget the first time I had (a few) people

tell me that because i recased my Falcon it was no longer a Falcon. Blah.

 

Adding an accelerator to an Atari ST doesn't change the fact that it is indeed, still an

Atari ST. Just a faster one. Deal with it. seriously.

Link to comment
Share on other sites

Oh crap, this isn't going to devolve into one of those, "If it ain't original Atari hardware, it's

not an Atari", discussions, is it?

 

I certainly hope not. That tired old argument that if you do anything to your Atari its not an

Atari any more is pathetic and lame.... I'll never forget the first time I had (a few) people

tell me that because i recased my Falcon it was no longer a Falcon. Blah.

 

Adding an accelerator to an Atari ST doesn't change the fact that it is indeed, still an

Atari ST. Just a faster one. Deal with it. seriously.

 

The TT isn't real! ;)

 

(sell me your TT cheap :) )

Link to comment
Share on other sites

You can not make STE from STf by replacing only CPU, not even if it is 100% faster. While it may help in scrolling - it will actually not, because code is made for STE specific HW. For DMA audio need lot of chips, complete new MMU, video chip, and so on ... I really don't see why people discuss things like this, while it is already solved - MIST. What can be accelerated little too.

Other thing is that so fast CPU needs changes/patches in TOS. And worse - most of Atari SW will work not, or with lot of issues on so fast CPU.

Link to comment
Share on other sites

You can not make STE from STf by replacing only CPU, not even if it is 100% faster. While it may help in scrolling - it will actually not, because code is made for STE specific HW. For DMA audio need lot of chips, complete new MMU, video chip, and so on ... I really don't see why people discuss things like this, while it is already solved - MIST. What can be accelerated little too.

Other thing is that so fast CPU needs changes/patches in TOS. And worse - most of Atari SW will work not, or with lot of issues on so fast CPU.

 

Yes - you are correct that you need to 'replace' more or augment more, and the MIST is one way to do it, although with somewhat CPU limited performance.

 

The Vampire Amiga core though DOES replace/extend these types of chips in the Amiga -- if you plug a Vampire accelerator into an Amiga 500 or 600, you get the additional graphics functions of the AGA chipset (introduced with the Amiga 4000), the enhanced 32-bit blitter, and HDMI audio output with capabilities beyond Paula. Vampire also goes one step further and adds 'SuperAGA' capability - adding a bitplane to the chipset design and extending graphics resolutions to 1920x1080.

 

There's no reason the Vampire couldn't do the same thing for an STf -- add the blitter and PCM Audio. In theory it could also add TT graphics as well. You could re-use some of the FPGA code from the MIST for the STE and implement on the Vampire; then you have a much faster CPU, the capability of the MIST, all on your ST..

 

Also, it's quite possible to have a FPGA cpu run at slower speeds or adjust timings to deal with those nasty TOS issues; there's already some discussion on the official Apollo CPU forums about this. The Vampire already handles the same issues on the Amiga as the Kickstart ROM is on the Vampire itself.

Link to comment
Share on other sites

Not quite sure why you'd stay with ST to STe when this should be able to more closely go from ST to TT and/or Falcon. Basically all the extra AGA stuff would be more equivalent of the jump from ST video modes to Falcon modes, since it went OCS/ECS to AGA.

 

Not to mention it'd could potentially (as mentioned in the thread on the apollo-core forum) that it could integrate the Super Videl into it. I think maybe a DSP on it would be the only thing preventing it from full Falcon capabilities.

Link to comment
Share on other sites

If I get how the Vampire 500/600 v2 works the HDMI is on the board so it ends up using the Amiga motherboard for very little (power, and peripherals).

I doubt you can have AGA graphics via the std video connector on the original motherboard given that Denise has only 4 pins (16 levels) for each R-G-B channel (http://www.ntrautanen.fi/computers/hardware/misc/denise.htm ).

So unless the Vampire requires soldering all over the place I believe that although technically you are getting an Amiga on steroids it could have been worthy to use an even bigger FPGA and replace the whole nine yards (MiST like which uses the same Cyclone III just smaller at 25K LEs vs the 40K LEs of the Vampire).

 

The V1 instead looks like it was more along the lines of extra mem (as other coproc boards), IDE connector, SD card and of course accelerated CPU all somewhat similar with the accelerators of old (which used real Motorola CPUs).

 

Lastly the Cyclone III used is no longer in production sigh, maybe that's the reason it can still be found relatively cheap in the market, even the software doesn't support it anymore (http://dl.altera.com/devices/ latest Quartus supporting the Cyclone III is 13.1 and it was released in 2013)

Link to comment
Share on other sites

Yes, more updates/improvenents you want it will be more wise to just drop old Atari completely, and make everything with FPGA and new chips.

Because it is simply simpler and more efficient to do so, instead interfacing with 30 years old chips and technology.

 

The Apollo/Vampire teams are working on a standalone version of 'Vampire' as well. They've been working on this 8 years, and their first versions intend to enhance the older computers (less work to do it this way) before moving onto the standalone. Ultimately the Vampire standalone should be a lot higher performance than the MIST.. Competition is a good thing :).

  • Like 1
Link to comment
Share on other sites

 

The Apollo/Vampire teams are working on a standalone version of 'Vampire' as well. They've been working on this 8 years, and their first versions intend to enhance the older computers (less work to do it this way) before moving onto the standalone. Ultimately the Vampire standalone should be a lot higher performance than the MIST.. Competition is a good thing :).

Actually we should not compare 2 different things. MIST is made with compatibility in mind, while Vampire with speed. I asked some questions on their board, and just as expected: there will be lot of problems with existing Atari ST, STE, Falcon SW. It executes all instructions of CPU range 68000-68060, but emulates none of it 100% accurately. Just something simple: stackframe - it uses 68040 type stackframe, so lot of 68000 code will fail miserably without patches. Not to mention cycle accuracy.

All in all, really no sense to go some ST expansion way. Make complete new HW, what will be something like Milan, Firebee - will need special TOS, special SW. Very small % of original Atari ST, Falcon, TT SW will run on it. I even don't see the point of whole thing, honestly. Fast CPU self makes not usable computer. Spending another 8 years on rest ?

  • Like 2
Link to comment
Share on other sites

Here's the foundation to this whole thread...

 

People want the Amiga and the ST to somehow be modern machines with a thriving, hardware and software community again. Essentially they want to return to the 1980's but with today's technology.

 

Who wouldn't want this? The 1980's was a golden era of computer development. It was a fun time and everyone asks the question "What if Atari had succeeded instead of IBM or Apple?" Its great fantasy material...and I'm guilty of wanting it too. However.. lets be practical.

 

We have a good retro computer platform with an existing software library. Some new software development happens, but not much. Anything that is made is typically in the area of games. Productivity software is better left to modern platforms for a myriad of reasons. But that being said, some ST productivity software could benefit from hardware improvements for very specific situations.

 

Thus anything that accelerates the ST is welcome provided it doesn't "break" the platform's ability to drive its existing software library. As such... I think "legit" accelerators that we see in the ST community today are the better way to go because they're centered around true compatibility rather than ultimate performance.

 

I'd love to see a Vampire ST if that somehow led to a resurgence of new ST productivity software development. But I just don't see that happening. Concentrate the Vampire for the Falcon and perhaps that could really happen. A Vampire Falcon actually has the foundation to be a modern desktop replacement...but the ST does not. The ST should stick with traditional accelerators.

  • Like 2
Link to comment
Share on other sites

I just watched the "World Premiere" video of the Vampire 500 on youtube. Its interesting that the designer's primary goal for the Vampire is to ultimately use it as the foundation for a "new" Amiga computer all together. Whether or not it actually sports the name "Amiga" is besides the point....but he wants to create an Amiga that can compete with modern day machines. The first step is providing these accelerators as "proof of concept" of what can be done with old hardware and the Amiga OS while being augmented by the Vampire. From there it will evolve into its own dedicated machine capable of running various cores as the user desires.

 

I think its ambitious...but no matter how active the Amiga (or Atari) community is... do you think it could be competitive enough to truly have a chance in the marketplace?

  • Like 1
Link to comment
Share on other sites

....

I think its ambitious...but no matter how active the Amiga (or Atari) community is... do you think it could be competitive enough to truly have a chance in the marketplace?

If all it took is an out of date Cyclone III 40K LE to be competitive .....

 

 

... don't get me wrong, it's a nice project and at the initial price of ~100US$ was attractive (before the guy got mad at the scalpers), as for being competitive on the modern market I just don't see it happening.

Link to comment
Share on other sites

Apple broke up with Motorola some years ago, and went on Intel CPUs (that was probably darkest day for Motorola), And it was not 680xx, but risk CPU PowerPC.

Why some think that can be competitive with architecture old over 35 years ?

I mean, coding 68000 is fun. Would be nice if things happened differently, but we are where we are now.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...