59 replies to this topic

[UNIXcoffee928]

Due to lack of interest this post has been removed.

Edited by UNIXcoffee928, Thu May 4, 2006 8:08 AM.

[warerat]

 UNIXcoffee928, on Wed Apr 26, 2006 8:16 PM, said:

The fact of the matter is that the Atari 8-Bit 6502 now has a backward compatible 32-Bit CPU available should have resulted in greater jaw-dropping, than the concept of putting the system on a PCI slot.

Don't think it's the CPU that's causing my jaw to drop...



Is this 2600khz?

[Heaven/TQA]

can the 6502 inside the 800 simply being replaced with this new cpu?

[UNIXcoffee928]

Due to lack of interest this post has been removed.

Edited by UNIXcoffee928, Thu May 4, 2006 7:53 AM.

[UNIXcoffee928]

Due to lack of interest this post has been removed.

Edited by UNIXcoffee928, Thu May 4, 2006 7:53 AM.

[ZylonBane]

 UNIXcoffee928, on Wed Apr 26, 2006 3:16 PM, said:

This could be the start of something very big, much as Linux was for OSs, this type of hardware could be revolutionary.

It could. But it won't. You seem to think everyone is just itching to build their own nonstandard computer if only the opportunity presented itself. Well... no. Computer hardware is and has always been a means to an end. It's all about the software, and investing in hardware that will run the widest range of it. It's why a kludgy architecture like the PC nonetheless managed to become the de facto standard computing platform.

People who are constantly blathering on about adding nonstandard new abilities to decades-old computers are hopelessly stuck in the past. The old 8-bits had a good run. They are what they are. Stop desecrating their grave.

[Rybags]

From a quick look at the schematics PDF, it has a max speed of 14 MHz.

One potential issue is DMA - but a workaround should be possible. In any case, the RAM would run at processor speed (or a multiple) and some sort of multiplexing system could be used to cater for DMA.

Duplicating the functions of ANTIC/GTIA would be possible, but relicating "features" like APAC mode, plus the vertical scrolling "bugs" might be a challenge.

If it was made as a PCI card, probably the best method would be to have it behave like a video capture card, ie - send bursts of screen data which would be integrated into a video overlay.

[RenegadeTK]

This sounds facinating. I am certainly going to watch this thread. I've got zero engineering skills, but I've been working with PCs and IT for several years now so if there is something I could do...let me know.

[Mathy]

Howdy folks

Me a great mind? Thanks!

Way back when even in the Netherlands magazines for the Apple II (two) were readily available, there was atleast one company that developed an upgrade that still has me puzzled. The Apple II GS has a 65816 built in. These upgrades replaced these 65816's with faster versions of the 6502. I have NO clue why they would want to do that, but they did. Why did they do so and would this have advantages for the 8 bit Atari scene? I've seen (on the net) versions of the 6502 that will do 166 MHz. That's a lot faster then the 65816's 14MHz.

greetings

Mathy (who's first Apple is an Imac G5)

[warerat]

Would any of this need to be backwards compatible, or is this just a totally new design?

[UNIXcoffee928]

Due to lack of interest this post has been removed.

Edited by UNIXcoffee928, Thu May 4, 2006 7:55 AM.

[DamageX]

Quote

You seem to think everyone is just itching to build their own nonstandard computer if only the opportunity presented itself. Well... no. Computer hardware is and has always been a means to an end.

This is true in general but there are lot of us hardware nerds who would be cranking out nonstandard computers left and right but for the time or skills.

Unfortunately, hardware nerds aren't all interested in the same kind of thing and lack of collaboration prevents most ideas from ever being realized. For instance, a multiprocessor Atari on a PCI card sounds like crazy talk IMO, but on the other hand, A8 upgrades like the XE videocard, 65816, and IDE are the kind of things I get excited about.

I thought the C-One was something that a lot of people would be interested in because it could offer a configuration to suit many different 8-bit computer fans. But I get the impression that somehow the popularity of these boards isn't quite we had hoped.

Part of the problem with expanding on the A8 is that it isn't very expandable and that Atari's own efforts at expanding it were rather poor (probably because the people who designed it in the beginning had left Atari, as well as price wars with Commodore). Because of this, probably most Atari users do not have expanded memory or a hard drive, and very few have a CPU upgrade. Wouldn't it be nice if you could take any Atari and just plug in a cartridge to get 256K or more of RAM, or whatever you want, instead of having to hack away at the insides? If it were so easy, then at least more users would have it and more software could take advantage of it.

Now back to the subject of a multiprocessor Atari in a PC... Just this week I heard that there is (or will soon be?) an FPGA that plugs into the CPU socket of a multiprocessor AMD Opteron board. I think that would be the way to go for your idea. Instead of having to design your own PCI card (which is also limitted by the speed of the PCI bus) you could use this already available hardware. The intended use for this was as a custom coprocessor to boost the performance of fancy-shmancy servers so it should be plenty powerful.

About those Apple II upgrades... as I understand there were a couple for the 8-bit apples that replaced the CPU with a 4MHz 65C02 and some cache RAM. The Apple IIC+ included something like this. Then for the GS there were accelerator boards with a 65C816 and 8KB~64KB of cache RAM running at speeds from 6MHz to 12MHz and beyond.

[UNIXcoffee928]

Due to lack of interest this post has been removed.

Edited by UNIXcoffee928, Thu May 4, 2006 7:55 AM.

[UNIXcoffee928]

Due to lack of interest this post has been removed.

Edited by UNIXcoffee928, Thu May 4, 2006 7:56 AM.

[PacMan]

I'm not sure what you are after here ?

Why do you want to integrate an atari into an fpga ? Your estimate of hacking it together in 3 months seems a little aggresive. I do this for a living and know that 3 months doesn't by you much these days. Maybe if you had a few people working full time. Wouldn't it be easier to take a one-chip ARM-9 and port a scaled down version of win800 or similar.

I started implementing the A800 in VHDL about 6 months ago and have managed to get about 50% of it working. I have designed it only from the available manuals and by running existing software (games mostly) to verify that the function blocks works. To verify 100% compatability I have then hooked a logic analyser to an old 800XL motherboard and compared the waveforms. This is actually what takes time... test and verification (as in any project).

But I totally agree with you, the 800 was by far the system that gave me the most pleasure, and it still does.

/P

[danwinslow]

Quote

Don't make me get negative. You won't like it much when I get negative.

Hehe, you don't know from negative. Look, you'd probably get a better response here if you dropped all the cutesy talk and the cheerleading. Somebody pops up around here every couple months all excited about some 'revolutionary' hardware thingy or game idea...and then usually nothing happens. This 32 bit 6502 sounds sorta interesting, but in itself it offers nothing new, as you can't take advantage of it by just directly replacing the CPU in an atari. That means that hardware glue will be required, clocks and ram and caches oh my!, and that stuff is hard and tedious to do. Then you'll have to rewrite the OS...And even then, what you get is not really anything different, its just faster and can address more memory ( which will still be banked somehow ). If all you want is faster then emulators can do that just fine.
If you can present some coherent scheme to actually *utilize* the processor in something that is still basically an Atari 8 bit, and is at least somewhat compatible with the software base, then by all means let us know.

[carmel_andrews]

In order for that to happen DanW, the following would need to be overhauled

New Atari mem. map, new memory sys. to rep. existing mem ands run highter then 2 mhz (i.e Sdr or DDr ram)

Expanded custom hardware (a new or revised antic/pokey/g/ctia) to access the new processor's new capabilities and also addit. data/address lines and run higher then 2mhz...same applies to the mainboard and how it comm's with the hardware/cpu etc and visa versa

Def. a new mainboard with add. data/address connections to PIA/Expanded antic/pokey etc and also EC/PBI ports (xl/xe only)

New or revise o/s with windews/t o s type front end, with ability to read in certain file formats from ST/amiga and PC/Mac

built in emu for old style xl/xe and 800 mode (for bk wards compat.

New o/s and front end to be compat. with hard drives and other ide/atapi devices (and scsi ) as well as USB/Firewire etc

An addit. bank switch location to compliement D301 or to expand d301 to 16/32 bit format (to bankswitch in more mem.)

in native 65xx mode only the first 64/128k is seen (in machines with sub 1meg mem.) in 65xx mode upto 4 megs can be used, the amount that can be accessed at one time is increased from 16 to 32 or 64k

In 65xxx mode (816 or 't32' mode), the first 1-2 meg is taken up by the system (incl. devices and device drivers, ram for custom hardware interupts and custom hardware like antic/pokey etc, disk sector buffering etc etc) the mem from there on in is paged in blocks of 4-16 megs (and upto 64meg blocks)all user programmable, expansion upto 3-4 gig poss (b4 bankswitching)

In respect to 65xxx mode the additional mem can be banks of 64k-1meg segments, like the current limit of 16k segement access

You can have different system config's and can hold upto 4-5 memory maps in memory (albeit in compressed form) and switchable between each system config and mem. maps without resetting

If you are using the emulated old atari mode, you will still be able to utilise the capabilities of the processor like speed and mem access as the emulated o/s and hardware will have been 're programmed' to the new processors abilities thru new or revised o/s patches

Any more thoughts

Edited by carmel_andrews, Thu Apr 27, 2006 5:46 AM.

[Rybags]

The reality of the situation is that just replacing the 6502 with a 14 MHz 16 or 32 bit descendant would probably involve re-engineering of most of the machine.

For starters, the existing hardware (ie RAM, ANTIC & co.) would probably fail if clocked too much over 2 MHz. In any case, ANTIC and GTIA need that 1.7~ MHz reference clock to generate a proper video signal.

I would speculate that it would be easier to just design an entire new motherboard, then populate it with the old POKEY, ANTIC and GTIA.

The CPU and RAM could run at full speed. Dividers would have to be used to give ANTIC and GTIA the 1.7~ MHz base clock they need to generate video, and for POKEY to work the same as in the XE.

Another problem is implementing ANTIC's DMA - extra interface hardware would be needed due to the difference in access speeds.

Finally, the ROMs would not be up to the task either - they would have to be replicated onto a faster access device (flashmem).

Edited by Rybags, Thu Apr 27, 2006 7:04 AM.

[UNIXcoffee928]

RAH 46 75 63 6B 69 6E 2C RAH! 04 07

Due to lack of interest this post has been removed.

Edited by UNIXcoffee928, Thu May 4, 2006 7:56 AM.

[mos6507]

The Terbium has been vaporware for ages and was discussed on the Commodore One mailing list.

If you really want to go down this route, write an enhanced Atari core for the Commodore One. There was talk about an Atari core from the start but nobody's ever started one as far as I know.

[drac030]

For ages? I can see that first rumours appeared on the net just 9 months ago.

It is true that there were some rumours about a vaporware WDC CPU trurly for ages, and there are even manuals available on the net (which seem fake), but this is the legendary 65C832 and not the 65T32. That last is at least mentioned on the WDC pages, and it wasn't last time I checked...

Edited by drac030, Thu Apr 27, 2006 11:06 AM.

[danwinslow]

Quote

All I can say is that the time is right. Lets do this!

Lol. Write when you get work, mkay?

[PacMan]

drac030->Definitly ages !,
I signed an NDA with WDC spring 2003 for their Terbium technology but I have not seen much from them since so it'll be interesting to see what they finally release.

/P

[Bryan]

This isn't a whole lot different than the people who always pop into the homebrew section with a game idea and insist someone else develop it.

Also, why isn't a 14MHz 65816 sufficient for upgrading an 8-bit? Who's going to want to write anything substantial for a 32-bit 800? And personally, I enjoy the obsolescence of the Atari.

/goes back to coding...

-Bry

[UNIXcoffee928]

This is rather funny.

Why on earth would I bother to tippity-type so many little characters, if I wasn’t planning on doing something?

It should be apparent to the astute readers that I have an extremely technical background, and a familiarization with the technologies necessary to achieve the goals that need to be met. Why presume I'm trying to 'insist someone else does it'? Other individuals may have been scared away from completing interesting ideas because of people who post such seemingly intimidating nonsense. I, however will state plainly that you're probably posers more than you are posters.

If you don't plan on offering something constructive, then go deride something else & don't waste my time, and the time of others who are interested in doing something with the project.

Anyway,

Since it appears that the link in the first Post was missed, why not request info from WDC?

Of course you could also join the Terbium mailing list:

You might want to Check the Wikipedia page on Bill Mensch:

Noting: As of 2006, Bill Mensch is still involved with design engineering at WDC in addition to his work as CEO. Among other technical tasks, he has written the upcoming Terbium processor family's data sheets and will be making the major RTL design decisions associated with that processor architecture.

Or, check the Wikipedia page for WDC


Have a Nice Day,

UNIXcoffee928