Was anyone else disappointed with the A1200?

[Dunny]

Originally Posted by hammer

Sales numbers matter unless you want Commodore to remain in the bankrupt timeline.

Why, are we going to change what happened by not enjoying Frontier?

[hammer]

Originally Posted by pandy71

I don't understand - if this topic is about A1200 vs SNES so why you constantly flooding thread with PC and TSENG - stay on topic, please.

SNES is an example. This topic is A1200 vs the the world.

Originally Posted by pandy71

[ Show youtube player ]

Unless you provide those documents then i'm right and you are wrong - check what is inside IBM document titled exactly same as yours for 8514 but about XGA:
http://bitsavers.trailing-edge.com/p...ence_Sep90.pdf

8514's Adapter Interface was recycled for XGA. LOL.

https://bitsavers.org/pdf/ibm/pc/IBM..._PS2_Jun88.pdf

From cited PDF, the title for part # 68X2279 is "Personal System/2 Display Adapter 8514/A Adapter Interface Application Developer's Guide".

IBM is asking for $25 for the part # 68X2279.

8514's compatibility is a cloner target e.g. https://ia804501.us.archive.org/28/i...sET4000VGA.pdf Page 27 for "8514/A Emulation Driver".

Originally Posted by pandy71

No registers there!

Cloners reverse-engineered and discovered registers and pushed IBM out of the PC market!

Originally Posted by pandy71

And other document (also seem to be unobtainum) covers AI i.e. software interface but no registers...

So unless you provide IBM document covering internal 8514 registers then you are providing nothing but misinformation - obviously you are not checking sources, instead purely relying on some random information's provided by google... We are aware how to use google...

From https://www.ardent-tool.com/video/85..._Wants_It.html

Microsoft (Redmond, WA) itself has gotten the 8514/A hardware specifications from IBM. Its sole purpose is to write Windows and Presentation Manager drivers-not drivers for application programs-to the 8514/A silicon because the performance simply wasn't acceptable when the drivers were written to the AI.

Microsoft has IBM's 8514/A hardware documentation.

Your argument wouldn't matter when there are many 8514/A clones with various price points.

By Windows 3.0's 1990 release, Windows 2D acceleration compatibility had a higher priority when compared to 8514/A e.g. S3 2D acceleration success in the PC market.

Originally Posted by pandy71

It was low enough - if you can afford for PS/2 50 or higher then you can easily buy VGA.

In real life, it's who you know. It was with my Dad's corporate contacts in Sydney CBD i.e. not a standard distribution sales channel. My ex-corporate PS/2 Model 55SX was traded in due to MCA's addon cards having a higher cost.

My family has at least two "32-bit" desktop computers from the early 1990s i.e. A3000/030 @ 25 Mhz (stock 1MB Chip RAM, 1 MB Fast RAM, obtained in H1 1992) and PS/2 Model 55SX (obtained Q4 1991). A3000/030's cost was offset by A500 Rev6A with 1 MB RAM, paid about $900 AUD after trade-in.

PS/2 Model 55SX was traded in for 386DX-33/ET4000 PC clone in Q4 1992.

With a single computer, IBM's 386 desktop PC products are within the budget i.e. $1500 AUD annual budget computer can build up for the big spend e.g. $3000 AUD or higher. With two "32-bit" computers to maintain, it's pushing the annual budget.

Parts of the $1500 annual budget on computers are spent on software.

PS/2 Model 55SX was primarily for "business use case" when it was loaded with Windows 3.0, Word Perfect, Ami Pro, Lotus 123, Harvard Graphics, and the old IBM Display Write. I tried to play (1990) Wing Commander on PS/2 Model 55SX and it was not ideal.
The Windows market came to be dominated by Microsoft PowerPoint in the expense of Harvard Graphics.

Originally Posted by pandy71

Now technical part to show how you manipulate with data.

IBM card is MCA so automatically market for this product is limited, then it has IBM quality (way superior) vs Asia clones and seem overall design is cleaner

IBM's quality is overrated when the primary CPU is "Intel Inside".

PS/2's metal case and chassis are strong and of good quality, but it's not needed for gaming PCs.

Originally Posted by pandy71

and probably faster than ET3000 but anyway this is completely unrelated to main topic so please stop flooding thread with random pieces about PC and TSENG.

IBM VGA is slow and it wouldn't beat A500's action 2D gaming capabilities.

With 1.5Ghz Cortex A53 PiStorm-Emu68 A500's HAM6 mode (19 fps) is faster than Athlon XP 2200+ (1.8Ghz) with original IBM VGA (8.6 fps).

25 fps is needed for PAL 50 hz.

IBM VGA doesn't impress me. It was SVGA cloners that improved PC gaming's fast VGA capability. IBM gave the strong 256 colors use case target.

Originally Posted by pandy71

Agree, let's stay on topic and no more Tseng

For A1200 vs the world, Tseng Labs can't be avoided.

https://dosdays.co.uk/topics/Manufac...tseng_labs.php
By 1991, according to IDC, Tseng Labs held a 25% market share in the total VGA market.

Originally Posted by pandy71

Already proved at least two times that you have no clue about yet you claiming things without justification, this was on e of your claims about ET4000 being compatible with 8514 (despite ET4000 having no hardware acceleration, not even hardware cursor) - any SVGA card can be 8514 compliant if you provide 8514AI driver but this not make it 8514 - native implementation of graphic functions will be probably faster than trying to implement 8514AI.

ET4000's Graphics Display Controller (GDC)

From https://ardent-tool.com/datasheets/Tseng_ET4000.pdf

Graphics Display Controller (GDC)

The GDC optimizes bit-mapped display memory data manipulation by assisting the CPU in the operation of displaying memory data-related functions. This includes rotate/mask/z-plane, with any of four boolean functions-in response to a single CPU write.

By putting basic bit map operations in high-speed hardware, the ET4000 dramatically increases graphics processing throughput over software-driven solutions.

The data manipulation capability implemented in the GDC is, however, applicable only for Plane systems and not for Linear Byte systems. This is because all the processing functions are designed to manipulate pixel data with one bit sourced from each plane.

For example, the color compare function allows four bits across four planes (one pixel) to be compared to a pre-defined color, thereby allowing eight pixels to be color-compared simultaneously by processing 32 bits of video data (one byte from each plane).


From Tseng_ET4000.pdf

Attribute Controller (ATC)
The internal Attribute Controller (ATC) provides flexible high-speed video shifting and attribute processing, designed for both text and graphics video display applications.

The ATC can process up to 16-bits of display data at the rate of 45MHz or 8-bit display data at a rate of 84MHz. In graphics modes, memory bits are reformatted into pixel color data in groups of 16, 8, 2, or 1 adjacent bits, translated through an internal 16-element color look-up table, and sent out serially to the video display.

Through this pixel mapper, the ATC supports "PLANE" (for 16 colors), "BYTE" (256 colors) and "WORD" (65,536 colors) oriented pixel structures.




System Priority Control (SPC)
The SPC's main task is to orchestrate the ET4000's internal resources requests including:

the FIFOs, Graphics Data Controller, Cache Controller, and RAM refresh. The system performance is based on two major factors: the ATC demand, i.e.; the display resolution and color; and the memory bandwidth, i.e.; the memory bus width and access time.

Other factors also can contribute to the overall performance. For example, the cache controller can be optimized for sequential access and CPU write operations.

The 16-bit CPU bus interface also results in faster data transfer, particularly in the plane graphics mode (a 16-bit CPU write = up to a 64-bit data transfer).


------------------
ET4000AX is more than IBM VGA and it's short of Blitter.

ET4000's Plane display mode is assist accelerated. LOL

ET4000W32 is released for "2D acceleration" which includes the Blitter.

The current Emu68's public RTG P96 driver is not 2D accelerated and it's still faster than "hardware accelerated" AGA.

[hammer]

Originally Posted by Dunny

Why, are we going to change what happened by not enjoying Frontier?

Are you against Commodore UK MD's argument for improved baseline AGA with minimal cost difference?

[Thorham]

Originally Posted by hammer

This topic is A1200 vs the the world.

That's definitely NOT this thread's topic It's in the thread title: Was anyone else disappointed with the A1200?

[hammer]

Originally Posted by Thorham

That's definitely NOT this thread's topic It's in the thread title: Was anyone else disappointed with the A1200?

The 1st post has two hardware comparisons against A1200.

[hammer]

Originally Posted by Thorham

Disappointed with the A1200? No, absolutely not. Best Amiga ever, made even better by the Blizzard 1230MK4 that I bought for it. Don't care if consoles from that time had more sprites/layers/etc. Try programming on those consoles

Peecees? Peecees of that time still sucked major donkey butt, of course, even if they were faster.

If I had to do it all over, I'd still choose an A1200 over a Pentium 90, even if given away for free

For the western markets, I preferred 68EC020-25 with Fast RAM as the baseline A1200/CD32 configuration.

Countries with weak currencies can have stock A1200.

[Bruce Abbott]

Originally Posted by hammer

Are you against Commodore UK MD's argument for improved baseline AGA with minimal cost difference?

I'm against further delays to add minor improvements.

I'm for an open-ended design that costs as little as possible and allows unlimited expansion with minimal redundancy.

For the western markets, I preferred 68EC020-25 with Fast RAM as the baseline A1200/CD32 configuration.

I preferred a cheaper price with FastRAM added via 3rd party expansions, like other low-end Amigas had.

Countries with weak currencies can have stock A1200.

It's not the currency that matters, but the affluence of users. In the UK cheap domestic home computers like the ZX81 allowed even 'poor' people to afford a home computer, and they could save more money by writing their own programs or even making their own hardware expansions. This created a vibrant hobbyist community that enrichened the experience.

In the US only wealthy people could afford the first home computers that came out there, but with four times the population there were enough 'rich' people to support the industry. The low-end was then filled with cheap game consoles. This resulted in a community of consumers who were little more than cash cows being milked by the industry.

If all you want to do is consume then just buy the latest product out there and consume it, but don't be surprised if you feel like something is missing.

[Dunny]

Originally Posted by hammer

Are you against Commodore UK MD's argument for improved baseline AGA with minimal cost difference?

How did you get there from my enjoying Frontier when I bought it on floppy (didn't have a hard disk in my 1200) back in 1994?

Which I did, btw. I really enjoyed it. Played it for months. You can't change that.

[AestheticDebris]

Originally Posted by Dunny

How did you get there from my enjoying Frontier when I bought it on floppy (didn't have a hard disk in my 1200) back in 1994?

Which I did, btw. I really enjoyed it. Played it for months. You can't change that.

I think it's possible to have enjoyed it and still be generally disappointed with the A1200.

Everyone I knew back in the day was disappointed with the A1200. The original Amiga had been an absolute step change in what was available and when a new Amiga with "advanced" graphics was announced, we were all expecting the same again, something that absolutely buried the competition and made Amiga the "must have" gaming platform it had previously been. What we got was something that felt little more than a minor upgrade, that lagged behind what the competition could already do.

Being disappointed doesn't necessarily mean the system was bad, but rather that it was too little and too late for what the market expected.

[OlafSch]

A1200 was a compromise. Commodore needed urgent to earn money so they could not afford to buy market share. RAM f.e. was extremely expensive at that time, also a faster processor. A1200 was created as a cheap entry system so I do not think that much more was economic possible. The problem was that it came much too late, A1200 should have been in market 1990 or even before to keep distance to the PC wortld. When it came out A1200 was ok and comparable to PCs in many senses but not more.

I was not disappointed at that time. I owned a A1200 and later a A4000 with graphic card and was a enthusiastic amiga supporter. A1200 was a big step compared to A500. But realistic it was too late already.

And if you look what happened with all the platforms and companies at that time, finally Commodore would not have survived propably. Only chance would have been to give up hardware development at all and concentrate on software and OS. If Commodore would have done less errors they would perhaps survived longer. But the chance to survive as a independent platform would have been unsecure and not very propable. Only Apple survived and even Apple would have not without money from Bill Gates.

[Dunny]

Originally Posted by AestheticDebris

I think it's possible to have enjoyed it and still be generally disappointed with the A1200.

Bravo, 100% agree. Not that I was disappointed - I came from an A1000 to the A1200 and I was literally in heaven even at stock configuration. There was so much more I could do with it even before I started on the games!

And there was no way in hell I could have afforded a PC. The 1200 cost two months wages as it was. And hell, as far as gaming was concerned, I messed around with 286s and 386s at work and jfc the games were terrible on those.

We had Zool on one. It was so bad.

[OlafSch]

@Dunny

+1

Compared to the existing hardware base with mostly A500, 1 MB RAM and two disc drives A1200, mostly with hard drives, were a big step ahead. I was not disappointed too at that time. For many amiga users changing to PC was no option. Looking from today at that time you can of course make a different judgement.

[Bruce Abbott]

Originally Posted by hammer

PS/2's metal case and chassis are strong and of good quality, but it's not needed for gaming PCs.

I remember when the clone manufacturers reduced the thickness of the metal in their cases to save money. Razor-sharp edges that would cut you, and cases so thin that they would bend under pressure causing cards to pop out of their slots. One of our suppliers resorted to hot-gluing VL bus cards into their slots. Cheap clone cases also had very cheaply made power supplies that were inefficient and unreliable.

But you are right, good quality wasn't needed for gaming. You were going to need a new PC anyway for the next game, so the machine only had to hang together for few months until you upgraded to a newer model. If it was flaky then all the more reason to upgrade!

The current Emu68's public RTG P96 driver is not 2D accelerated and it's still faster than "hardware accelerated" AGA.

I doubt that. PiStorm's ChipRAM interface is quite slow, ~3.7MB/s max compared to 7MB/s on a stock A1200 or well-designed 68k accelerator card. Are you sure that Picasso 96 can produce the same effects as 'hardware accelerated' AGA (sprites, dual playfield, copperlist, HAM8)?

[Thomas Richter]

Originally Posted by Bruce Abbott

Are you sure that Picasso 96 can produce the same effects as 'hardware accelerated' AGA (sprites, dual playfield, copperlist, HAM8)?

No, but does AGA offer hardware overlays, true-color and 3D acceleration?


Fact is, it's a horse of a different color. The reason why standard hardware evolved as it did is that the use cases changed, and they were optimized to such use cases. Nobody needs a copper and HAM if true-color is available, nobody needs sprites if a fast accelerator can move graphics around, and nobody needs dual playfield if hardware overlay is available. The way how the the average Amiga game worked and was programmed is significantly different from the way how the average PC (or raspi) game works today. Tasks for which you needed dedicated hardware back then (sprites) are no longer relevant as the accelerator can move graphical objects around much faster without sprite limitations (such as limited width and limited priorities), so they are not really practical today anymore. Other units were added, such as 3D accelerators, that were completely out of reach at Amiga times since hardware complexity was much too high.

[Dunny]

Originally Posted by Bruce Abbott

I doubt that. PiStorm's ChipRAM interface is quite slow, ~3.7MB/s max compared to 7MB/s on a stock A1200 or well-designed 68k accelerator card. Are you sure that Picasso 96 can produce the same effects as 'hardware accelerated' AGA (sprites, dual playfield, copperlist, HAM8)?

Chip reads are 3.5mb/sec but writes are the full 7mb/sec according to bustest.

I'm not much of an expert but I'm not sure that p96 uses ChipRAM for anything much at all? I'm running a 1080p workbench display and have the full 2MB of ChipRAM free.

And the CPU driving the display (Emu68's 68040) is clocked at around 2.2Ghz - which although variable in speed due to JIT is certainly faster than any physical 68k CPU for any operation.

[Thomas Richter]

Originally Posted by Dunny

I'm not much of an expert but I'm not sure that p96 uses ChipRAM for anything much at all?

It doesn't. Unless you open a planar mode, where the bitplanes are, in fact, allocated from chip mem to allow applications to use the native blitter to operate on such planes. Planar is really only a backwards compatibility feature, and not very practical.

[pixie]

Originally Posted by Thomas Richter

No, but does AGA offer hardware overlays, true-color and 3D acceleration?

Fact is, it's a horse of a different color. The reason why standard hardware evolved as it did is that the use cases changed, and they were optimized to such use cases. Nobody needs a copper and HAM if true-color is available
(...)

I guess doing copper tricks like gradients are trivial in rtg, still it would be nice having them on rtg in a 1:1 fashion. For example I have not seen the possibility to do magic copper in rtg despite being just a 'fixed' gradient in a chroma colour.

[Dunny]

Originally Posted by pixie

I guess doing copper tricks like gradients are trivial in rtg, still it would be nice having them on rtg in a 1:1 fashion. For example I have not seen the possibility to do magic copper in rtg despite being just a 'fixed' gradient in a chroma colour.

Probably very difficult to do when your display surface is 32bpp and has no actual paletted colour indexed pixels like 8bpp and below do. Which pixels do you swap for the copper-style gradient?

[pandy71]

Originally Posted by hammer

SNES is an example. This topic is A1200 vs the the world.

This topic is about A1200 disappointment not about superiority of 486 over 386 - SNES is example of some HW but compared to Amiga not PC vs PC. Stay on topic.

Originally Posted by hammer

8514's Adapter Interface was recycled for XGA. LOL.

LOLipop - XGA was designed after 8514 and it closed gap between VGA and 8514 providing graphic acceleration from IBM together with VGA compatibility. Additionally offered accelerated 64k colors screens.
But... But this topic is about A1200 disappointment not about superiority of XGA over 8514.

Originally Posted by hammer

https://bitsavers.org/pdf/ibm/pc/IBM..._PS2_Jun88.pdf

From cited PDF, the title for part # 68X2279 is "Personal System/2 Display Adapter 8514/A Adapter Interface Application Developer's Guide".

IBM is asking for $25 for the part # 68X2279.

8514's compatibility is a cloner target e.g. https://ia804501.us.archive.org/28/i...sET4000VGA.pdf Page 27 for "8514/A Emulation Driver".

All above is SOFTWARE - why the heck you willing to emulate 8514AI (i.e. 8514 API) if you can probably provide higher performance by directly perform operations using CPU for your unaccelerated graphic HW.
And clearly you are providing misinformation as IBM officially not provided 8514 register documentation (as they goal was to use dedicated API i.e. 8514AI) - by using reverse engineering number of 8514 HW compatible clones was created anyway.

Once again - this topic is about A1200 disappointment not about correcting your misleading arguments about for example ET4000 having accordingly to you HW acceleration when Tseng staying completely opposite. You simply don't see difference between software blitter and HW blitter.

Originally Posted by hammer

Cloners reverse-engineered and discovered registers and pushed IBM out of the PC market!

From https://www.ardent-tool.com/video/85..._Wants_It.html

Microsoft (Redmond, WA) itself has gotten the 8514/A hardware specifications from IBM. Its sole purpose is to write Windows and Presentation Manager drivers-not drivers for application programs-to the 8514/A silicon because the performance simply wasn't acceptable when the drivers were written to the AI.

Microsoft has IBM's 8514/A hardware documentation.

Your argument wouldn't matter when there are many 8514/A clones with various price points.

By Windows 3.0's 1990 release, Windows 2D acceleration compatibility had a higher priority when compared to 8514/A e.g. S3 2D acceleration success in the PC market.



In real life, it's who you know. It was with my Dad's corporate contacts in Sydney CBD i.e. not a standard distribution sales channel. My ex-corporate PS/2 Model 55SX was traded in due to MCA's addon cards having a higher cost.

My family has at least two "32-bit" desktop computers from the early 1990s i.e. A3000/030 @ 25 Mhz (stock 1MB Chip RAM, 1 MB Fast RAM, obtained in H1 1992) and PS/2 Model 55SX (obtained Q4 1991). A3000/030's cost was offset by A500 Rev6A with 1 MB RAM, paid about $900 AUD after trade-in.

PS/2 Model 55SX was traded in for 386DX-33/ET4000 PC clone in Q4 1992.

With a single computer, IBM's 386 desktop PC products are within the budget i.e. $1500 AUD annual budget computer can build up for the big spend e.g. $3000 AUD or higher. With two "32-bit" computers to maintain, it's pushing the annual budget.

Parts of the $1500 annual budget on computers are spent on software.

PS/2 Model 55SX was primarily for "business use case" when it was loaded with Windows 3.0, Word Perfect, Ami Pro, Lotus 123, Harvard Graphics, and the old IBM Display Write. I tried to play (1990) Wing Commander on PS/2 Model 55SX and it was not ideal.
The Windows market came to be dominated by Microsoft PowerPoint in the expense of Harvard Graphics.



IBM's quality is overrated when the primary CPU is "Intel Inside".

PS/2's metal case and chassis are strong and of good quality, but it's not needed for gaming PCs.


IBM VGA is slow and it wouldn't beat A500's action 2D gaming capabilities.

With 1.5Ghz Cortex A53 PiStorm-Emu68 A500's HAM6 mode (19 fps) is faster than Athlon XP 2200+ (1.8Ghz) with original IBM VGA (8.6 fps).

25 fps is needed for PAL 50 hz.

IBM VGA doesn't impress me. It was SVGA cloners that improved PC gaming's fast VGA capability. IBM gave the strong 256 colors use case target.

Thanks for sharing this amusing history of your family - it was nice lecture but we agreed earlier that this topic is about A1200 disappointment not PC (and other mostly Japanese platforms) amusement.
So please stay on topic.

Btw we had agreement that VGA (and EGA) offered some HW features providing possibility to close gap in terms of performance.
If you wish to learn more on this i can recommend this interesting lecture:

https://fabiensanglard.net/another_w...DOS/index.html

Originally Posted by hammer

For A1200 vs the world, Tseng Labs can't be avoided.

Oh, how's that? Commodore planned to replace Alice+Lisa with Tseng product?

Obviously your goal is try to derail this thread by not staying on topic and flooding it with completely irrelevant and misleading information's about Tseng products - from your perspective ET4000 is almost same like Et4000w32 etc even if both are completely different products sharing only (partially) name - you are quoting some information's but you don't understand it as it is irrelevant from topic perspective. Please stay on topic!

[pandy71]

Originally Posted by Dunny

Probably very difficult to do when your display surface is 32bpp and has no actual paletted colour indexed pixels like 8bpp and below do. Which pixels do you swap for the copper-style gradient?

You simply do things different way, your code operate on color values not on registers index. You need more CPU cycles but still this at the composition level - modern graphic use HW compositors so you can have various planes in memory and at the composition stage they just mixed with preprogrammed weight (alfa) so in theory you can have more planes and they can be mixed at various stages (using blitter and HW plane composer) - in 3D HW this can be even more extended - beyond traditional 2D plane concept.
Probably Copper tricks are easier to be done nowadays at the pixel level than before on line level.