Jump to content
IGNORED

Is the C64 too different to A8 to ever have meaningful comparison?


oky2000

Recommended Posts

uhm..ok.. But games look and play better on the a8. And the TIA and derivatives is faster than the vic-ii in c64.

 

I wanted to express that, I THINK, wanted to improve the 2600 chipset.

Or to put it another way: they (or the management) didn't knew better.

 

Just speculating here.

However, isn't it a fact that the AMIGA is to the A8 the same as the A8 to the 2600?

ANTIC generates interupts (DLIs) so the CPU can change registers.

The Copper does change registers directly.

POKEY has IRQs based on the frquency register (AUDFx). In the ISR the CPU can shove sample data into AUDCx and play samples at varriying frequencies.

PAULA does it directly.

AFAIK the AMIGA even has PMG-like sprites (although they werent used much?)

 

So it all boils down to Jay Miner's ideas of how the chips should accomplish the needed tasks.

Or maybe he was stuck on hiis way of thinking.

For example, why hasn't the AMIGA real (C64 or NES style) sprites?

 

 

What I do not know (and really would like to) is when were "sprites" invented.

I assume there is research paper somwhere which proposes "a layer of moveable objects above a bitmap".

Link to comment
Share on other sites

1) The Commodores had split up quite some time ago (Lionel Richie sings solo)... no more support for The Commodores by Lionel Richie means better/higher support for the A8 of course... ;-)

Are the Ataris still going though? I used to quite like listening to them...

 

Why else would the C64 be that colour?

 

(My first C64 originally belonged to a small team of chain smoking software developers, when i dismantled it to repair the wear and tear i found the board had yellow patches on it!!)

My dads mate used to code on his C64 whilst chainsmoking cigars. I used to be able to sort disks by smell.

  • Like 2
Link to comment
Share on other sites

2600, A8, Amiga - the heritage is evident when you look at the overall philosophy of graphics generation.

 

But there is a quantum leap in each case, actually the 2600 to A8 one is probably the most significant, in that the 2600 has a downright primitive playfield generation methodology (20 bits repeated or mirrored, PF or background colour only).

 

Looking at each, you can see that the successor is like a wish list was presented to programmers and each new machine fulfilled those wishes.

I remember going from 6502 to 68000 programming and thinking much the same of the instruction set. 68K is very much a fulfilment of what most 6502 programmers wanted or coded macros/subroutines to achieve.

 

On Amiga there had to be compromises in that only so much DMA bandwidth is available for playfield, sprites and sound, remembering that cycles still had to be left over for the CPU, Refresh, Copper and disk I/O.

 

Sure, like A8, C64 and many other old computers the DMA usage was kind of subtle in that a lot of cycles were still left over - it would have been perfectly OK in each case to have 100% saturation of cycle usage devoted to the chipset (esp graphics) during the visible display to allow for better graphics/sound generation.

 

I suppose one thing forgotten is Amiga had it's roots a good couple of years or so before actually hitting the market and was initially to be a console.

Also we should compare to the contemporaries of the time of release and expected competition, realistically nothing touched it until the Megadrive/Genesis and SNes consoles came out, it took PC and Mac a number of years to surpass it.

 

Of course with Amiga there's the age old argument that the chipset improvements were too little and sometimes too late. 16 bit sound was something sorely missing which realistically should have appeared in the 1200 if not earlier.

Link to comment
Share on other sites

POKEY don't get no respect!

 

All the SID this SPRITE that. The POKEY was the greatest multifunction chip ever put in a computer. TI came out with an IBM multifunction chip that included IDE, Floppy, Printer, and Serial ports. I consider the TI chip mundane in that it came out so much later and was less of a building block for a functional computer hack.

 

Consider this: If any of the hardware guys here were asked to design a simple four chip computer with some glue logic in say 1988, what would they do to give the most bang for the buck? It most CERTAINLY would not be a SID. Look what POKEY brings to the table, first and foremost is serial capabilities: ~112k BAUD right out of the box and everything else down to a couple of hundred baud. 6551? Peshaw! 19,200 BAUD in big steps.

 

Now think about adding a keyboard. Even today people are doing the AVR which is a full blown micro to translate an IBM style keyboard which has a full micro in it! POKEY allows you to do just about any level of keyboard with interrupt capability. POKEY adds a 64 key matrix and if you want, can even do stuff like handle controllers on a 5200.

 

Want some A/D? Not voltage level which would be nice, but charge pot. There are even programs that use it for digitizing audio and playing it back. 8 channels!

 

The audio of course! 4 channels.

 

The timers! Elegant solution to just about all your needs. The CTC used in other computers? Lots of pins, not a lot of extra functionality over POKEY.

 

*IF* someone asked me today for a lowest common denominator computer, it would probably be a micro<6502 preferred>, 32k RAM, 32k EPROM<or smaller to allow for expansion>, POKEY. Matter of fact, if someone asked me for just about anything needed the above capabilities irrespective of processor<if they weren't built into the processor> first thing I would add to my blank sheet of paper would be a POKEY. I am surprised there haven't been a lot of designers that have done this. Then again, the number of home brew computers vs. people hacking stuff like Arduino or Raspberry Pi is pretty small.

  • Like 3
Link to comment
Share on other sites

Ok, if someone were to say something like "Dual POKEY", would ricortes completely blow a gasket? :grin: (That's all the working interrupts, serial, and keyboard, PLUS four working sound channels all at the same time.)

 

It would make one heck of electric keyboard/synthesizer! A real piano only has 88 keys. You would have another 40 keys for changing things on the fly like vibrato and sustain. The 16 pots could be used to translate knobs to channel volume/fuzz like a mixer. Of course with two POKEY, you could add MIDI in and MIDI out.

 

I see you get it, using POKEY as a building block. :) An 8 voice, 16 channel A/D, 88 keys +40 user defined, stereo or mix, one heck of a machine and only takes 5 chips. I forgot to add hardware random number generator to my first list so two with a dual POKEY. Just one heck of chip for 1979. I don't think it has ever been equaled. We are just starting to catch up with some of the latest ARM/AVR chips but for them, everything like keyboard scan and sound generation would have to be done in software. Read and debounce a key, sound blips.

Link to comment
Share on other sites

Pokey has a weakness with keyboards in that the scanning doesn't allow multiple keypresses other than Shift & Control + another. Plus there's no known way to scan individual key states.

 

Also there's the 2 multiplexing ICs but of course they are mainly to reduce the pin count on Pokey itself.

 

As for multifunction capabilities available 10 years later - plenty of scope. Even a few years after A8 came out Antic and GTIA were able to be combined into a single die.

 

Going by Moores Law alone, it should have been feasible to do the entire Atari chipset in a single package by 1982 or so. It's not necessarily a case of being able to fit components into a single die, the pin count in some cases will make doing individual chips more economical.

 

SID was impressive in that some ways the audio generation is as simple as the method used in Pokey. The Phase Accumulator method allows both the frequency and waveform generation to be handled by much the same set of components.

 

Putting serial and keyboard in the sound chip or otherwise really doesn't matter much in the grand scheme of things. To an extent it becomes a weakness because we lose a couple of voices when doing SIO. But as a benefit we gain the signature I/O sound which practically no other computer has.

Edited by Rybags
Link to comment
Share on other sites

Valid points, but firstist with the mostist and still not equaled. Trivial to get it to function as a 112k BAUD serial port in existing system and with a different clock, could easily hit spot on 115k BAUD. If the only thing you wanted to do was add a serial chip to a system you have very limited choices. POKEY's contemporaries were things like the buggy 8250 and 6551. I *think* Zilog made a DUART, I forgot the numbers. The 6850 was a good serial chip, but it was only a serial chip. Nothing was faster for years, 6X the speed of a 6551 used in an MIO is nothing to sneeze at. Fewer lines for handshaking of course. Meh! With 8 pots you could do some pretty good handshaking at least on input. Still a 115k BAUD MIO would have been a nice upgrade to a 19.2k BAUD one. All that ICD would have had to do is wire in a POKEY vs. the 6551.

 

From a system standpoint, I don't think you could go wrong with POKEY as it's basis. Just gives you so many options. For argument sake, lets say you wanted to implement some flavor of 68xx. If you want it to be a serial terminal interface to real world, POKEY will handle it. If you want it to have its own keyboard, you can wire up a non encoded keyboard matrix and do your own system or use an existing keyboard. Anything from an Atari keyboard to a ZX81 could be used. I'm pretty sure you could even decode a C64 keyboard with it. This is beautiful flexibility,

 

There's a huge number of peripherals, everything from disk drives to printers, that will interface directly with the POKEY. We aren't just talking Atari peripherals, anything RS232 with level conversion.

 

If you wanted roll your own cassette interface, you could do Kansas City Standard. It would be a little tougher reading the tapes w/o a decoder chip but everything back then did just that. Obtuse the way I am trying to make the point. Trivial to encode tapes, same difficulty as everyone else decoding them.

 

For that matter, SIO2USB and SIO2SD would be vigorous choices for the hypothetical system if it was currently being designed. One measure of success is devices are still being designed that use it.

 

POKEY does everything I could conceivably need in a small system except video. Not enough pins left on it to add an ANTIC and GTIA! :)

 

Remember this is a somewhat rigged test. Four chips, most bang. You absolutely need [processor, RAM, ROM]. What forth chip can you add to give you the most functionality? I say people will be hard pressed to come up with a better chip to add then a POKEY and this is 34 years after its introduction. All the other computers contemporary with the Atari like the Kim, Challenger, Super Elf, AIM-65, TRS80, they wish they would have had a POKEY. :-D

Link to comment
Share on other sites

Pokey serial is versatile in that the array of possible speeds is vastly greater than most.

 

But there is a shortfall in that the mode of operation is fixed, ie you must have start/stop bits. Of course there's the option to do bit-banging but very CPU intensive and you can essentially do the same on a PIA port bit.

 

TED excels in it's graphical ability but falls well short in most of the other areas it covers. Sound is weak, if the 2600 had 10 bit frequency control then it'd be well ahead of Ted. Basic latched type parallel I/O I doubt is very expensive in terms of chip area, really it's the pin count that usually means a PIA/VIA/CIA is used in other machines.

Link to comment
Share on other sites

TED actually has a better saturation level on it's colours than both A8 and C64 - I get the feeling there is variation built in as the darker ones aren't oversaturated and the brighter ones aren't washed out like A8.

 

I guess I could one day test the theory, I've got a real Plus/4 here and oscillosope.

Link to comment
Share on other sites

 

Certainly a serious contender for 4'th chip. IMHO, there would be applications and it should have been a popular chip to hack. Problem with it is a lack of I/O. You end up with an extra 8 pins, five years later, no attachment to peripherals other then unencoded keyboard<or switches> as input and video + sound as outputs. A four chip system based on it would make a good standalone video game. A matter of personal preference, I would rather have I/O in a system then video. If you add a serial chip capable of 112k BAUD, the argument could be made a POKEY based system could add a 6545 for 80 column video. Best computer with five chips is an interesting design challenge. Kind of wide open at the five chip stage.

 

You probably could do some I/O, but it would be annoying slow. Something like FSK on audio out and add a ~PLL to decode cassette audio in with the lock appearing as a keyboard press. Interesting academic challenge adding storage I/O to something not designed for it. I also decided to consider having a cart port on a four chip system as adding a 5th chip! :)

 

I actually did something like that back in the day. Bit banged on an Atari J/S port to transfer files to an ST J/S port. ST had an odd port, behaved more like a keypad with respect to reading then a digital I/O port. I got terrible through put, much less then 300 BAUD.

Link to comment
Share on other sites

For cassette, I really liked what they did in the Color Computer. A DAC / ADC was used with the CPU to read the cassette data. That system ended up being very robust and fast! The Atari and POKEY did this in blocks. Blocks were kind of nice because there was some time for processing things as data came in, but they weren't too nice just due to how slow that cassette system was.

 

Can this be sped up? Just wondering.

 

Anyway, the CoCo just dumped it all as one big bit stream. Start it, load it, done, stop. Blocks could be done easily, and it's a software only process using the CPU like that. Seems to me a tape could be used very well, and on the CoCo was, including file names and types.

 

I second the I/O priority. In the end, that proves much more useful. But, I must say having at least 80 column capability native made a big impact on how a machine was taken back then. External 80 columns would basically mean it's not used much, and even at that high baud rate, might not work well for many things that onboard video would...

 

Interesting discussion!

 

Seems to me, one could make a similar argument for sound happening via serial too.

Link to comment
Share on other sites

Lol! I was really missing this A8 vs C64.

Capitain to the Earth now I just taked control of this site.

Keep them going but with respect.

 

And respect means that no more than 20 of 25posts per page molesting the poor C64, VicII and Sid isn't allowed.

Outer than this and this Thread will self destroy in the next posts if no post have "C64" in the text!

Keep them going but with respect, O.K.?

Link to comment
Share on other sites

The Atari and POKEY did this in blocks. Blocks were kind of nice because there was some time for processing things as data came in, but they weren't too nice just due to how slow that cassette system was.

 

Can this be sped up? Just wondering.

Not much. Os++, the operating system that comes with Atari++, will have in its next release a "turbo" tape version that should work natively on hardware. However, the trouble here really is that parts of the hardware to decode the stream is in the tape deck, and not in the main computer system itself, which somehow limits the available speedups.

 

What Os++ does is that it doubles the size of the blocks (which reduces the gap overhead) and it also slightly speeds up the baud rate, still to be within the limits of the hardware. This is probably as far as it gets. Does it help? Not much.

Link to comment
Share on other sites

Mentioned before but I had an AsmEd image which loaded the entire 8K Rom in a single block with higher bitrate.

 

Realistically the blocksize increase has limitations in that the buffer size required becomes a bigger hinderance than the benefit of eliminating IRGs. Also tape stretch could become a bigger problem with larger blocks.

 

But for something that does burst loading without buffering there should be no problem going with 1-2K block size.

Link to comment
Share on other sites

Lol! I was really missing this A8 vs C64.

Capitain to the Earth now I just taked control of this site.

Keep them going but with respect.

 

And respect means that no more than 20 of 25posts per page molesting the poor C64, VicII and Sid isn't allowed.

Outer than this and this Thread will self destroy in the next posts if no post have "C64" in the text!

Keep them going but with respect, O.K.?

Luckly we don't have to fight like the UK C64 users with the Spectrum owners, C64 vs ZX Spectrum ouch....

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...