Atari v Commodore

[atariksi]

>i was keeping the hardware aspects apart and comparing the differences between moving and colouring of sprites without involving other aspects like software sprites. i'll be honest, i'm using the term "keep it simple" with a heavy degree of sarcasm now because it only seems to apply under certain circumstances... comparing four players making two objects to two hardware sprites is keeping it simple, actually thinking about how much code required to move those players around not so...

 

I meant the algorithm can be complex-- the hardware aspects are simpler if they are separated. You have been using "keep it simple" in a disparaging way as if someone is restricting you. We were talking multicolored sprites (including you)-- in Atari a multicolor sprite is by definition two single color sprites. The Atari has a better implementation of the multicolor sprite. You have more multicolor sprites on the C64. Motion in the Y-axis is faster with multiple PMBases-- yeah with some restrictions depending on useage.

 

>Well, i could be wrong and i've not been awake long... but from what i remember that doesn't allow independent vertical movement of the players so in most real world situations it's not exactly going to hold it's own, in fact it doesn't even allow for animation of the objects (which my method does) and chews through a fair block of RAM in the process...

 

It does-- it depends on usage. If you cross the two multicolor sprites in the same zone you have to resort back to memory move or adjust the PMBase tables for high probability. For a single 7+ color sprite or where sprites are in zones, it works great. Your better off arguing it's subjective like you did with the palette.

[drac030]

Just a few corrections: What you call a "program" is a "display list" -> just a table of display zones.

 

Interestingly, C-64 people, as long as it fits their argumentation, call a display list a 6502 program, that feds VIC with data:

 

http://en.wikipedia.org/wiki/Display_list

 

Unfortunately, this is not the case here. The "display list" on Atari is a term describing a series of ANTIC op-codes, a program.

Oh my, the worst wikipedia article I have read in a long time. Quite a lot of bullshit there.

 

Oh my, but, as long it seemed to fit your argumentation, you cited wikipedia as well, see the post 335. No inconvenient "bullshit" there, I presume?

 

So since ANTIC doesn't read a program, the pointer to the list is no "program counter".

Unfortunately, the pointer to the display list isn't identical to the ANTIC program counter. Could you learn some basics about atari internals before starting trolling here?

I know the ANTIC and GTIA pretty well.

 

If you mix the DL pointer with the ANTIC PC counter, then you do not.

 

And a processor processes data, like "calculating something", "rendering polygons" or atleast "rendering bobs".

I can't find "rendering bobs" here: http://en.wikipedia.org/wiki/Microprocessor

"A microprocessor incorporates most or all of the functions of a central processing unit (CPU) on a single integrated circuit (IC)."

 

And the "rendering bobs" was referring to GPU/VPU, not to microprocessors.

 

I told you above, that ANTIC is not a GPU, so what's the point of requiring "rendering bobs" from it afterwards? (I am pretty sure that you will skip this question in your answer).

 

And sure, ANTIC incorporates "most of" the functions of a central processing unit on a single integrated circuit: it is a single integrated circuit, it has a program counter, it has separate instruction list, and it processes data. Of course, it is a specialized microprocessor (but in no way a video microprocessor: it is a video microprocessor only in the sense that the circuit it is controlling is a video interface controller), but it is still a microprocessor.

 

While the VIC-II is only an I/O chip. ChipSelect line etc. Like 6520. And so you in no way can claim the C-64 to be superior in design here For that, however, better go to a ZX Spectrum forum - you can legitimately claim that a C-64 displays more colours, and, of course, that the ZX Spectrum's ULA doesn't generate violet; and, IIRC, it doesn't generate brown either.

Edited November 26, 2008 by drac030

[atariksi]

please code a sprite multiplexor with free screen positioning for each sprite and look how much you can display... player/missles are good for Mule etc (and time bars) but a pain for fast action games esp. involving repositioning on screen...

 

another detail which gets lost... the 4 missles are encoded in 1 byte so moving them independently means additional code, too again... (masking/or'ing data).

 

For items requiring small number of sprites horizontally, it's not so bad. I was more interesting to know what was your limitation with the fonts exactly in that game you mentioned. Where the conflict exactly occurred which made it undoable via DLI.

[Fröhn]

Oh my, but, as long it seemed to fit your argumentation, you cited wikipedia as well, see the post 335. No inconvenient "bullshit" there, I presume?

Wikipedia article quality differs. Seems that the display list article contains fanboyism combined with half- and no-knowledge. But I must admit that I stopped reading at the point where C64 FLI and Amiga HAM were put into the same bin because obviously the writer thought that Amiga HAM was some kind of software driven mode which is pure bullshit.

 

If you mix the DL pointer with the ANTIC PC counter, then you do not.

DLBASE is not the same as DL pointer (you may also call it DL counter if it makes you feel better).

 

"A microprocessor incorporates most or all of the functions of a central processing unit (CPU) on a single integrated circuit (IC)."

And where exactly does ANTIC do that? Where's the ALU for example?

 

I told you above, that ANTIC is not a GPU, so what's the point of requiring "rendering bobs" from it afterwards? (I am pretty sure that you will skip this question in your answer).

But you said it was a VPU. VPU is an alternative term for GPU.

 

While the VIC-II is only an I/O chip. ChipSelect line etc. Like 6520.

Huh? What are you trying to say by that? The CS is there to write VIC-II registers. ANTIC has a register set written by the CPU too in case you forgot that.

 

And so you in no way can claim the C-64 to be superior in design here For that, however, better go to a ZX Spectrum forum - you can legitimately claim that a C-64 displays more colours, and, of course, that the ZX Spectrum's ULA doesn't generate violet; and, IIRC, it doesn't generate brown either.

Seriously I consider a 1 IC design which is doing similar things superior to the more expensive 2 IC design. Especially if it can do better 2D games

[frogstar_robot]

Why "wrong"? Before Tramiel bought Atari, Atari didn't want Miners Amiga project. That's why he left and build the Amiga company. Or do you believe that only because he worked at Atari for a while they kinda own him?

 

I believe no such thing. I believe Oswald to be a fanboy who sees anything with an Atari badge on it as bad and primitive and anything with a Commodore badge on it as flawlessly good. It is nonetheless a fact that the Amiga tech could have been used in an Atari project and if that had happened then Oswald would be here today trashing it. I've indicated in another post that it was probably for the best that it wasn't and I've indicated in yet another post that that Warner was the impetus behind Atari's best and brightest leaving them.

[frogstar_robot]

I tend to just think of Antic/GTIA as being the display processor. I guess you could connect a GTIA directly to a 6502 in some way - but it wouldn't be that much better than the original 2600

 

Isn't it possible to drive the GTIA directly from the 6502 ala the 2600 with a "display kernal"? Not that it would happen much.

[frogstar_robot]

good luck finding c64 wingnuts who are trying to claim a bit of fame from a non c= system. they dont need to.

 

That is strictly your take on it and not at my intention. As I mentioned before but I'll state it again: I respect and admire Jay Miner's genius irregardless of who his employer happened to be at any given time. Even the TIA from the 2600 has a certain elegance and quite the advance over the PONG-on-a-chip it was designed to replace.

 

I've also stated I have no brand loyalty. Atari was an innovator during Bushnell's tenure but not so much during the Warner and Tramiel years. Were I fanboy, I'd have no criticisms for them. I have no problem conceding that Commodore did a better job of marketing and supporting their products during those times. For that matter, I liked many of their later products better than what Atari had. Even the Falcon and TTs didn't hold much of a candle to the Amiga line. But even the mighty Commodore had a few flops and they too were ground under the Wintel juggernaut.

 

The mutually incestuous history of Atari and Commodore in the mid eighties is what it is. And strictly speaking, the Amiga was mostly developed before Commodore bought it.

Edited November 26, 2008 by frogstar_robot

[Allas]

4 - BALLBLAZER

 

Atari screenshots

 

This is definitive a good example how the Atari can manage better games again his counterpart. This time, C64 did a great port, but it looks simple against a 128 color palette, a faster CPU, a better scrolling system. Simply, is a unreachable quality for a C64 machine. Is necessary to view the motion and exploding effects to take count of all the show.

 

For sure it can be done several games in Atari with this quality, it's a pity Lucasfilm team lost in the old days with only a few projects.

 

C64 screenshots

Edited November 26, 2008 by Allas

[TMR]

Atari computers were always more stylish, that's all you need.

 

Why do i find the phrase "style over content" springing to mind...? =-)

Actually Content and style! Those C64's were Ugggg leeee!

 

i'm sorry, you seem to be offering subjective opinions that are outside the bounds of this discussion. Have a nice day.

 

 

Talking about yourself?

 

Significantly higher numbers of the "ugggg leeee" machines were sold so either people really didn't give a crap how a machine looked (and lets face it, only a mother could love the 400) or, if how a machine was styled was important, they must have disagreed strongly with both of you since the "stylish" Ataris were out-sold rather heavily.

 

Have a nicer day.

 

i already was, ta.

[Tezz]

Isn't it possible to drive the GTIA directly from the 6502 ala the 2600 with a "display kernal"? Not that it would happen much.

Yes it is

[Crazyace]

Driving GTIA directly only allows you to set players and missiles. The playfield is only supplied by the AN0-AN2 circuitry ( and Sync as well ) so there isn't actually as much functionality exposed as the original TIA

[TMR]

I meant the algorithm can be complex-- the hardware aspects are simpler if they are separated. You have been using "keep it simple" in a disparaging way as if someone is restricting you. We were talking multicolored sprites (including you)-- in Atari a multicolor sprite is by definition two single color sprites. The Atari has a better implementation of the multicolor sprite.

 

Yes, i do feel it's being used as a one-sided restriction which is why i've been sarcastic when using the term. If you use two sprites to render one object on the Atari it simply isn't a fair comparison to insist that the C64 has to "keep it simple" and isn't allowed to use the exact same technique; the better implementation of multicolour sprites is the hardware that can do it either way depending on what the programmer wants.

 

It does-- it depends on usage. If you cross the two multicolor sprites in the same zone you have to resort back to memory move or adjust the PMBase tables for high probability. For a single 7+ color sprite or where sprites are in zones, it works great. Your better off arguing it's subjective like you did with the palette.

 

i can't see how it's subjective as to which is a better design; one takes CPU power, RAM overheads and restricts motion of objects depending on the code-driven solution used or the other that merely involves setting Y registers. If we were really keeping it simple, the code would be kept out of it and players can't actually move vertically at all without the code...

[Crazyace]

I dont know how anyone can argue that the Atari sprites are better than the C64 sprites - it's just not true..

 

( Setting double width on C64 - so that the res is the same )

 

C64 - 8 sprites, each 24 pixels wide

A8 - 4 sprites, each 8+2 pixels wide or 5 sprites, each 8 pixels wide

 

....there's no comparision - the C64 wins the sprite award with ease

[Heaven/TQA]

please code a sprite multiplexor with free screen positioning for each sprite and look how much you can display... player/missles are good for Mule etc (and time bars) but a pain for fast action games esp. involving repositioning on screen...

 

another detail which gets lost... the 4 missles are encoded in 1 byte so moving them independently means additional code, too again... (masking/or'ing data).

 

For items requiring small number of sprites horizontally, it's not so bad. I was more interesting to know what was your limitation with the fonts exactly in that game you mentioned. Where the conflict exactly occurred which made it undoable via DLI.

 

aehm... having 8x16 multicolour sprites would requiere to take all 2 players to get 1 enemy...and I have more than 4 in one scanline... and I am using players for weapon and armor overlays...

[frogstar_robot]

I cannot say either is a bad machine, there problems were more from their corporations. Both companies had great products and should have been exploited, the same way the IBM PC was.

 

The IBM name counted for a lot of business mindshare right off the bat. It also helped that the PC was entirely COTS except for the BIOS. Prior to the PC businesses looked at micros and the companies making them as hippy enclaves. This was perhaps less true of a company named "Commodore Business Machines" but then nobody ever got fired for buying IBM as they used to say then. The reason why PCs ultimately won and Macs these days are really cleverly designed PCs is that they commodtized computing. Ataris, TIs, Commodores, Apple IIs, and all the rest were proprietary architectures that mostly didn't interoperate in terms of either software or hardware and only the Apples were cloned with varying degrees of legality. Once Compaq showed that it was possible to create a legal "IBM Compatible" BIOS that opened the door for lots of companies to create mutually interoperable machines.

 

The cleverness of things like the Amiga and even the ST (even an ST made a typical clone PC look pretty sick until the VGA adapters and Adlib cards started coming out) didn't stand a chance against the developing ecosystem of third-party development that commodity PCs made possible. Apple hung on quite a bit a longer by paying closer attention to ergonomics but all Apples made now use components that aren't out of place in a PC and many are running software from that world as well.

 

Were Commodore a bit more adroit and Atari a LOT more adroit then the computing history of the eighties and early nineties might be a bit different. The end result was still inevitable and in time the stodgy PCs caught up to and surpassed those competitors to the point that they can emulate them in a window and still have the horsepower to get a rant posted on AtariAge.

[Bryan]

I don't have any problem saying that for most games, the C64 has the clear advantage.

 

The Atari chipset was designed when there were NO chipsets out there that could do a full alpha screen plus sprites, not to mention 4-voice sound. Atari was the pioneer in this area and they clobbered the VIC-20 which was released at about the same time as the 400/800. When VIC II and SID were being developed, the Atari was the machine they needed to beat.

 

Atari's crime was letting their engineers leave and their hardware stagnate.

 

-Bry

[Shannon]

Sticking to the originating message...

 

8-bit miner 2049'er > C64 version..

 

And the sound/music in Jumpman on the C64 sounds like a buzzsaw..

Edited November 27, 2008 by Shannon

[ThomSW]

I wanted to write sth really harsh but better I leave alone this already openned can of worms...

[Oswald]

Say anything you want, that wouldn't change the bare facts: antic is a microprocessor, it has program counter, own instruction set, fetches and processes data according to this program and independently of what 6502 is doing, can perform jumps and nops and so on. Of course, its program is primitive (it can't do arithmetic calculations or memory writes, but that's what's 6502 for), but it still is a microprocessor.

 

the fact is that a microprocessor=cpu=> cpu operation has 4 steps at minimum: "The program is represented by a series of numbers that are kept in some kind of computer memory. There are four steps that nearly all von Neumann CPUs use in their operation: fetch, decode, execute, and writeback." antic is missing the execute, and writeback features. antic is a statemachine with some cpu-esqueish features (jmp). the display list is used to change its inner state, which in turn affects the graphical output. however it cannot execute anything that gets near to a program, a program would need atleast the writeback feature, if then style instructions, etc.

 

(from http://en.wikipedia.org/wiki/Central_proce...#CPU_operation)

[Oswald]

Let's put something "new" into the round

 

You cannot imagine a good looking 3D action game on the A8 with 8-12 fps?

Well, I can. The only rule would be to have the screendata on a cartridge...

 

wolf clones even in a8 demos are under 25fps. guess the speed on a 4x bigger screen... +add the sprite handling, AI, yaddayadda, you're at 2-4fps at best.

 

Please don't mixup the software modes that are used for most demos with the hardware mode where Antic reads every 2nd line from it's own buffer.

Softwaremodes take CPU speed... Hardwaremodes give CPU speed back, and using players overlayed by the playfield means no DMA cycles. Only some DLIs are necessary for actuating the pm graphics where needed.

Have a look at Space Harrier and how fast it runs. A Wolf 3D would need lesser DLIs ... And, yes, with double scanline modes you also can have ingame digi voices!

 

Aint most 4x4 atari doom clones using a HW mode where antic reads gfx only at every 4th line? So a mode compared to this, which reads at every 2nd line will slow down the cpu alone. Also even without deep programming knowledge you should realize that calculating 4x as much pixels will take about 4x as much time (4x4 vs 2x2 mode), and that comes without adding game logics, AI, sprites, etc.

 

Space Harrier is wonderfully done, but it runs on a charscreen requiring updating 16x less bytes than the mode you are proposing for this game.

[Oswald]

Why "wrong"? Before Tramiel bought Atari, Atari didn't want Miners Amiga project. That's why he left and build the Amiga company. Or do you believe that only because he worked at Atari for a while they kinda own him?

 

I believe no such thing. I believe Oswald to be a fanboy who sees anything with an Atari badge on it as bad and primitive and anything with a Commodore badge on it as flawlessly good. It is nonetheless a fact that the Amiga tech could have been used in an Atari project and if that had happened then Oswald would be here today trashing it. I've indicated in another post that it was probably for the best that it wasn't and I've indicated in yet another post that that Warner was the impetus behind Atari's best and brightest leaving them.

 

I dont have such views, nor I denied amiga tech could have been used in an atari project. furthermore if my grandma was yellow and had wheels she would have been a taxi.

[Oswald]

I've also stated I have no brand loyalty.

 

guess thats why you go into personal attacks when I dont share the views of some atari "wingnuts".

 

I have no problem conceding that Commodore did a better job of marketing and supporting their products during those times.

 

sure you have no problem. thats the only reason why they sold stuff better than atari, not because they had better machines, right?

 

 

And strictly speaking, the Amiga was mostly developed before Commodore bought it.

 

depends on how you define mostly. if you mean the main design of the HW I agree, but certainly many times the work went into it in its c= era, than before. ECS, AGA, OS, etc.

[emkay]

Space Harrier is wonderfully done, but it runs on a charscreen requiring updating 16x less bytes than the mode you are proposing for this game.

 

Depending on the fine movement of the objects, it would be odd to use a charmode. Compare to the charmode usage of the C64 version, which moves in the resulting charmode rasters, you see the difference.

Same thing with Yoomp! .... No charmode, no trick... just cpu power and Antic co processing... and, well, some GTIA PM graphics

[Crazyace]

That's the thing missing with the 8 bit atari's. The Amiga had revisions to it's basic chipset to improve things. The ST had hardware revisions as well ( MegaST , STe, falcon ) - even the c64 had the c128

But all of the Atari revisions never changed anything with Antic/GTIA

[Oswald]

That's the thing missing with the 8 bit atari's. The Amiga had revisions to it's basic chipset to improve things. The ST had hardware revisions as well ( MegaST , STe, falcon ) - even the c64 had the c128

But all of the Atari revisions never changed anything with Antic/GTIA

 

 

the c128 is really just a c64 with 2mhz mode, a 80 column video chip, 128kram with banking logics, and a useless (clocked too slow) z80 cpu. damn commodore for not gong for a 65816 and a VICIII.. they have crippled the whole c64 successor idea with the damned 99.9 compatibility.