Was anyone else disappointed with the A1200?

[hammer]

Originally Posted by Bruce Abbott

I was told this RTG P96 driver is 'not 2D accelerated'. You seem to be talking about something else, perhaps a graphics processor in the Pi's SoC? In which case who knows what tricks it's using.

For Emu68, the current public P96 RTG driver for Emu68's Pi SoC IGP is not accelerated.

[ Show youtube player ]

Michal Schulz: Please note there is *NO ACCELERATION* in use here. Neither for decoding nor for the graphics driver. It is all mastered purely by Emu68 m68k JIT.


It's just CPU's brute force like the old-school PC VGA style. The dumb framebuffer is fast.


[ Show youtube player ]
Some 2D acceleration with the WIP P96 driver was shown on YouTube. It's better with some 2D acceleration.

[hammer]

Originally Posted by Thomas Richter

The last time I was impressed is when I got hands on the latest AMD Ryzen. Also, the latest Raspi is quite impressive for its form factor. There is no "impressive" hardware anymore at my home - I simply haven't found a use for it, all I have is good enough for my needs, and if I would buy something new with similar performance, it would be in many aspects worse. No DVD anymore, no usable keyboard anymore, no legacy connectors anymore, instead a full bag of adapters for USB-to-LAN, USB-to-USB, USB-to-VGA, USB-to-HDMI, HDMI-to-VGA... Thanks, but no thanks. I want to plug in my LAN cable, my DVI monitor, my USB-2 mouse, and I want to play my DVDs, and not pile up electronics.

I have a Framework laptop with slotted adapter cards for XYZ to USB-C/TB.

https://www.notebookcheck.net/Framew....552116.0.html

[sandruzzo]

Like I said the real deal with Aga was the commodore itself! Even with AAA they would have fallen!

[OlafSch]

@Thomas

At that time I owned both, a A1200 I bought first and a A4000 with Picasso 2 I bought around 1994. I think both were different systems. Most amiga user at that time were just playing games, only a minority used it productive. And even less were used professional. The A1200 was designed as a replacement for A500 and as a entry system competing with cheap entry PCs. For that it was OK and at least to me no disappointment. The problem more was the system was too late and no longer better than standard PCs. For productive users like you A1200 was not interesting of course. One big problem was, games were developed on PC (MSDOS) and later ported to Amiga. The architecture on PC was different and more relying on resources, expecially ram and processor. Additional the problem with missing chunky at amiga. A1200 was ok but would have needed expecially more horsepower to stay competitive or games using the advantages of amiga. But additional efforts were not profitable for the software developers because of the smaller market.

I yesterday did some experiments in amigaforever with using HAM8-Mode in a graphic app. I could not seriously work with such strange and low resolutions. As long you are only playing games it is ok but for doing something more serious it is impossible to me. As soon you have used something modern you never look back. But that is all from today, we talk about young people with limited finances around 1992-1994. A different world.

[Dunny]

Originally Posted by Arne

Boot times were great in the Amiga/8bit/16bit days, then got abysmal with W95/98, and now they're at least tolerable.

Bear in mind that they're only tolerable now because everyone has SSDs. Without that, they're bloody awful again. In fact, overall system performance falls through the floor if you don't have a hard disk that is essentially just RAM.

That's how bad things are right now.

[OlafSch]

Originally Posted by Dunny

Bear in mind that they're only tolerable now because everyone has SSDs. Without that, they're bloody awful again. In fact, overall system performance falls through the floor if you don't have a hard disk that is essentially just RAM.

That's how bad things are right now.

from a technical view might be... but users look at at it how it feels to them. And if they can work comfortable with it they do not care if it could have been better or different.

[pandy71]

Originally Posted by hammer

Read https://eab.abime.net/showpost.php?p...postcount=3516
You started PC addon vs PC addon. PC vs PC is NOT my argument!

Nope, if anyone started something with PC it was you not me - you obviously have obsession with Tseng products same like with some Japanese computers (NEC if i recall correctly).
So instead blaming others for not staying on topic blame your self only.
Various VGA cards performance in Doom on PIC is completely out of this topic.

Originally Posted by hammer

What's the intent of Vampire and PiStorm, again? A resurrection via retro with a forward road map direction.

Well main purpose is to pass MC68060 prices and borders - max you can get from MC68060 is 133MHz - as Motorola replaced 68k family line by CF then people seek affordable way to provide more power to Amiga computer line.

Originally Posted by hammer

TheA500mini has reached mainstream brick-and-mortar stores in Australia.

Go for some PC Thin Client and have way better A500mini - this is all i wish to wrote about A500mini.

Originally Posted by hammer

https://medium.com/@scottdaniel_2089...-ff5c90ddde85#
The Rise of Retro Gaming: Nostalgia in the Digital Age

Well not me - i'm not retro gamer - Amiga is for me way more interesting as personal computer than game station.

Originally Posted by hammer

Less than two years. May 1985 to April 1987 is 23 months. Don't exaggerate.

PC has a strong 256-color use case.

Amiga's 2D acceleration becomes de-acceleration with fast 32-bit 68K.

Well - my memory serve something about Moore law and tick tock and similar things - 18 months? double transistors? something like this?

This is whole generation in computers, more memory, higher speed and/or lower cost.

256 for every computer is "strong use case" - limiting factors are RAM size and RAM bandwidth.

Originally Posted by hammer

You're a fool to compare Amiga's 2D acceleration with the SVGA chipsets that enable Amiga's RTG.

lol - first, MC68000 AFAIK has no no RTG support so RTG is possible only in 68020+ machines (i.e. less than 30% of all Amiga machines) - they are mostly beneficial from native Amiga HW acceleration, also some very important aspect is developers knowledge - know how, they was exposed to HW acceleration, this is very important for next generation of developers. I always say that this is one of the most underestimated values of Amiga - people learned how to use blitter, HW line draw, etc.
Popularity of Amiga created developers skills, attracted people wiling to learn graphic HW - this was beneficial for whole PC industry.

Originally Posted by hammer

Lew Eggebrecht: "We can't make that decision right now - it's something we'll have to look at but in that time frame, even in the low end, every machine is likely to have a DSP. It's a cost thing - although the AT&T chip itself is only $20 to $30 or so. AT&T has a number of lower cost options, as well, that are designed more specifically to go on the motherboard


DSP3210 has a $20 to $30 price range and it's in the game plan, but Commodore runs out of time and money.

Lew Eggebrech is open for Nintendo-style officially supported DSP for ALL Amigas including low-end Amigas.

It's too bad Eggebrech was not in the hot seat earlier in the 1988 time frame. Lew Eggebrech is better than Bill Sydnes.

DSP32 is a new math-based object manipulator.

And ? How this is related to topic?

Originally Posted by hammer

8514's AI wasn't enough for Microsoft and IBM gave hardware documentation for 8514.

Once again - IBM officially never released 8514 register documentation in public - if Microsoft received such documentation doesn't change fact that it was not known to all developers and as such your next point is valid.

Originally Posted by hammer

8514 does nothing for PC DOS gaming. 8514 compatibility is nothing when S3 didn't even bother implementing it.

S3 implemented similar interface to 8514 - but it was also not known in public so from overall development perspective it could never be a standard.
There was no industry standard for graphic acceleration until WinG and later DirectX, in parallel OpenGL and similar interfaces (like Glide)

Originally Posted by hammer

With Windows 3.0's 1990 release, I didn't care about 8514 compatibility. The only credit for IBM is the strong 256-color use case as the baseline clone target.

Well, 8514 never exposed frame buffer (i.e. video buffer) to CPU as such frame buffer will not work for 8514 obviously - this is exactly same approach as NEC uPD7220 and many similar graphic accelerators where frame buffer is never available in CPU address space - if you wish to set pixel then you need to program graphic accelerator.
Windows for sure was not best 8514 target.

Originally Posted by hammer

You started the PC add-ons vs PC add-ons debate. Read https://eab.abime.net/showpost.php?p...postcount=3516

Nope, not me - you started this and it was way earlier than https://eab.abime.net/showthread.php...78#post1678578 .

Originally Posted by hammer

8514 does nothing for PC DOS gaming.

Suprise suprise - why 8514 was unpopular for PC gaming?

Originally Posted by hammer

It covers your "professional" 2D acceleration argument.

8514 does nothing for PC DOS gaming.

I use 8514 as example that even on PC some players realized necessity to introduce HW graphic acceleration and IBM was natural leader, foundation of standards like VGA standard for example, market closely followed IBM.
By closing 8514 specification and making it MCA attached IBM automatically killed product mass market popularity but anyway triggered some idea on PC - this idea started to grow slowly.

Originally Posted by hammer

That's your argument. I didn't advocate ET4000's inclusion into AGA.

Nope - you are constantly arguing about various VGA card performance in some various benchmarks and games like Doom - so we have endless loop of various pictures and numbers how good was ET4000 in Doom.
No Doom in Amiga, no ET4000 in Amiga, no VGA in Amiga so this is completely out of this topic.

Originally Posted by hammer

My ET4000xxx with market share example is to show the competitive market from the gaming PC's side and it's against the "PC has crappy Tridents" argument.

Once again - topic is about A1200 disappointment no about superiority of Tseng ET4000 over Trident 8900C - obviously you are constantly going off topic. And Trident was more popular because its lower price so obviously not everyone was interested in Doom performance.
Nowadays large part of PC market are embedded graphic controllers (IMHO they are dominating market).

Originally Posted by hammer

I advocated for improved math power for A1200/CD32 as per Lew Eggebrech's interview.

Great - so i assume with your strong support we can expect new Amiga models with DSP on board - as you and Lew will convince CBM management.

Originally Posted by hammer

Somebody argued with most PCs have crappy Trident SVGA. I posted Tseng Labs' market share and the size of PC's annual unit sales.

Well... Tseng for sure never was majority of PC market and this closing topic. Your analysis about Tseng market share are simply limited.
Tseng was not so popular as you think.

Originally Posted by hammer

I posted Intel's 486 and Pentium shipment report to shoot down the "most PCs are XT/AT crap with monochrome display" argument.

Once again - you are assuming wrongly that new PC base is bigger than old one when it was exactly opposite - for single sold 486 machine there was many 386 machines.... PC in those times was not replaced after year or two.

Originally Posted by hammer

Can't you read? ET4000 can assist the CPU with certain workloads. VGA has some accelerated features since emulated VGA is slow i.e. ask Rendition, Inc.

I can read so this is my question - provide detailed list of operations where ET4000 assist CPU in a way not present on other VGA chips.
And CPU assistance is not HW acceleration where HW replace CPU utilization.

Originally Posted by hammer

Your argument wouldn't matter when AGA's 2D acceleration is not fast i.e. A1200's gimped 68EC020 has 49% of A1200's Blitter.

Well - as blitter performance was same as in ICS/OCS/ECS then for sure fast CPU can be faster than not improved HW.

Originally Posted by hammer

You're a fool to compare AGA's 2D (de)acceleration against SVGA chipsets that powered Amiga's RTG cards.

You are fool saying this - if HW was slow then you could use CPU - nobody force you to use slower HW. This should be clear also for you.

Originally Posted by hammer

I haven't forgotten OS patches for fast CPU bilts on my A1200 with 8 MB Fast RAM card, and later with TF1260. The OS patches for fast CPU bilts were applied to my A3000(68030@25 Mhz with static ZIP Fast RAM)'s OS.

"Deaccelerator" says Hi.

If you have fast CPU then it is valid, if you have slow CPU then it is not valid. This is very simple. Not everyone had 68060 with lot of RAM in his A1200

Originally Posted by hammer

https://www.xfree86.org/4.0.2/Status24.html


"OTI087 (the latter with some acceleration) is provided by the XF86_SVGA server and the oak driver."

Instead third party information check vendor datasheet for "acceleration":
http://www.vgamuseum.info/index.php/...bb9d36ef57beaa
What kind of acceleration (bitblit? line draw? fill? etc) support OAK OTI087?
There is not even HW cursor (sprite) there...

[dreadnought]

Originally Posted by Dunny

Bear in mind that they're only tolerable now because everyone has SSDs. Without that, they're bloody awful again. In fact, overall system performance falls through the floor if you don't have a hard disk that is essentially just RAM.

That's how bad things are right now.

Whoa, much hyperbole

Seriously though, are you guys sure this is only because Windows is so unmitigably awful, and has nothing to do with the small matter of modern O-systems being somewhat more complex than the +30 years old ones? (and that's even without the absurdity of comparing them to 8-bit)

Snark aside, I am genuinely curious. I have no doubt that assorted Windows bloat contributes to that but even the mighty Linux takes some time to boot, so...

[robinsonb5]

Originally Posted by dreadnought

Seriously though, are you guys sure this is only because Windows is so unmitigably awful, and has nothing to do with the small matter of modern O-systems being somewhat more complex than the +30 years old ones?

It absolutely has to do with modern operating systems being more complex - but that runaway complexity is itself a huge problem, and one with no easily attainable solution in sight.

And yes, modern Linux is just as bad - booting it from spinning rust is very much like booting an Amiga while its boot volume is validating.

[Bruce Abbott]

Originally Posted by Thorham

Current and recent GPUs are quite interesting.

How so?

[Thorham]

Originally Posted by Bruce Abbott

How so?

Ever tried writing shaders?

[hammer]

Originally Posted by pandy71

Nope, if anyone started something with PC it was you not me - you obviously have obsession with Tseng products same like with some Japanese computers (NEC if i recall correctly).

That's a false narrative.

I'll post it again:
[ Show youtube player ]

Doom (low details) on 386DX-40 with 128K cache
Tseng ET4000 ISA = 26.751 fps
Trident 8900CL ISA = 23.0088 fps
WD90C32 = 26.838 fps (Diamond Speedstar 24X)

That's three SVGA graphics cards.

NEC's PC-98 wasn't IBM PC compatible just as different 68K based desktop computer platforms like Atari ST, Apple Mac, Amiga, Sharp 68000 and 'etc'.

Originally Posted by pandy71

So instead blaming others for not staying on topic blame your self only.
Various VGA cards performance in Doom on PIC is completely out of this topic.

1. My argument wouldn't change when SNES with SuperFX2 had a Doom port in 1995.

A1200 would need a fast 68030 CPU accelerator card for Doom's near full screen with low details and floor/ceiling textures.

68EC020-25 could be acceptable at a minimal cost increase and without floor/ceiling textures.

2. Lew Eggebrecht has low-cost DSP32 in the roadmap for low and high-end Amigas. I support the unified Lew Eggebrecht/Dave Haynie/Mike Sinz's DSP32 inclusion.

Originally Posted by pandy71

Well main purpose is to pass MC68060 prices and borders - max you can get from MC68060 is 133MHz - as Motorola replaced 68k family line by CF then people seek affordable way to provide more power to Amiga computer line.

ColdFire is rubbish for Amiga's legacy game use case.

Originally Posted by pandy71

Go for some PC Thin Client and have way better A500mini - this is all i wish to wrote about A500mini.

That's not an "out of the box" experience.

Originally Posted by pandy71

Well not me - i'm not retro gamer - Amiga is for me way more interesting as personal computer than game station.

This topic is about games.

When the Amiga's "power without the price" gaming reputation was damaged, Amiga's non-game market also declined.

The Video Toaster niche market will not sustain Commodore.

Amiga's "power without the price" gaming was the entry point for Amiga's non-gaming use case.

Amithlon's "we don't care about games" is f__k you classic Amiga games legacy.

"AmigaOS X86" is like Windows NT on non-X86 i.e. a fish out of water.

Originally Posted by pandy71

Well - my memory serve something about Moore law and tick tock and similar things - 18 months? double transistors? something like this?

486DX was released in 1989 and pushed 386DX into the lower product tier.

A similar situation when Pentium 60/66 was released in 1993 and pushed 486s into a lower-tier product stack.

The cycle repeats as Intel continues to release higher-clocked P5 Pentium and Pentium Pro's 1995 release.

From the wholesale price front, 68040 was tracking 486DX prices. Motorola didn't factor in the 486 cloners.

The PC market had superior big chip component distribution channels.

68060's 1994 wholesale price is competitive, but 040 socket accelerator boards are not cheap.

There's an economic of scale issue with 3rd party Amiga CPU accelerators. Each Amiga accelerator card design is unique for each Amiga model. There's no common compute module for A500/A1500/A2000/A1200 when compared to Raspberry Pi's small board computers(SBC). The low-risk additional cost is the gateway PiStorm adapter board for each Amiga model.

Raspberry Pi's innovation is its' non-PC form factor with low cost. Unlike PowerAmiga or other PowerPC attempts, Raspberry Pi didn't bother copying PC ATX standards. Raspberry Pi redefined low-cost barebone small computer form factor.

Imagine, a common 68060 compute module that can be used from A500/A1500/A2000/A1200 and standalone Linux68K. 68060 accelerator card wouldn't be bound by A1200's smaller market size but by the combined A1200/A500's larger market size. The common 68060 compute module would be spread to other 68K based systems. Replace 68060 with another big-endian CPU family. That's a hindsight after RPi form factor innovation.

Originally Posted by pandy71

lol - first, MC68000 AFAIK has no no RTG support so RTG is possible only in 68020+ machines (i.e. less than 30% of all Amiga machines) - they are mostly beneficial from native Amiga HW acceleration, also some very important aspect is developers knowledge - know how, they was exposed to HW acceleration, this is very important for next generation of developers. I always say that this is one of the most underestimated values of Amiga - people learned how to use blitter, HW line draw, etc.
Popularity of Amiga created developers skills, attracted people wiling to learn graphic HW - this was beneficial for whole PC industry.

There's very little point in talking about developing the next generation developers when there's an exponential cost beyond Commodore's mass-produced A500, CD32, and A1200 models.

Commodore UK MD argued for upgraded CD32 into the mass production process. Lew Eggebrecht's direction seems similar. Commodore ran out of time and money.

Economies of scale matters and the Amiga platform lost this mass production capability when Commodore went bust.

The Amiga platform has access to RPi's economies of scale on the CPU and RTG side. The next step is to clone the AGA (or SAGA) retro platform at the lowest price possible. ARM-based SOC is the future of the Amiga platform, but retro gaming is the entry point.

Skills development is outside the scope of this topic.

Originally Posted by pandy71

And ? How this is related to topic?

Commodore UK MD argued for upgraded CD32 into the mass production process. Lew Eggebrecht's direction seems similar. Commodore ran out of time and money.

Originally Posted by pandy71

Once again - IBM officially never released 8514 register documentation in public - if Microsoft received such documentation doesn't change fact that it was not known to all developers and as such your next point is valid.

For B2B relationships, IBM had released 8514 hardware documentation to MS. Windows 8514 driver doesn't fully rely on 8514's AI since it's insufficient.

Microsoft takes the risk when 8514's underlying hardware changes or IBM provides its XGA driver.

IBM doesn't guarantee 8514's underlying hardware will remain in successive generations.

For the PC, VGA is the last generation for hit-the-metal before shifting to Windows 3.1 and OS/2 or Windows 95/NT. Duke Nukem 3D used VBE.

Sony has other ideas with PS4's hit-the-metal GCN/RDNA 2.

VGA is only used for safety mode.

Originally Posted by pandy71

S3 implemented similar interface to 8514 - but it was also not known in public so from overall development perspective it could never be a standard.
There was no industry standard for graphic acceleration until WinG and later DirectX, in parallel OpenGL and similar interfaces (like Glide)

Windows up to version 3.x and early versions of OS/2, the only interface available for graphics was the Graphics Device Interface(GDI). This defines a large number of 2D graphics primitives, from simple line-drawing to font rendering. Graphics drivers implemented the GDI Device Driver Interface (DDI), with primitives matching the GDI primitives.

Early accelerators were sold as "Windows accelerators" and provided better versions of various GDI DDI primitives, e.g. line-drawing, rectangles, polyfills, font rendering in some cases (e.g. hardware caches), hardware mouse cursors, etc. This involved moving primitives, or parts of primitives, from the drivers into hardware.

Later, the acceleration focused on primitives which were more useful to accelerate, in particular for Windows games. The main desktop-relevant optimization was hardware BitBlt (e.g. WinG which provided fast BitBlt and it's not necessarily hardware-accelerated). The leaked WinDoom port used WinG/Win32S and this game still works on Windows 11.

Windows 95 introduced its mini-drivers based on a DIB engine: instead of implementing everything, drivers could delegate functions to the generic DIB engine (cite Windows 95 DDK documents). Acceleration efforts switched to DirectX, in Direct2D and Direct3D, with primitives designed with optimization.

Comparing GDI and Direct2D hardware acceleration gives a good overview.

Originally Posted by pandy71

Well, 8514 never exposed frame buffer (i.e. video buffer) to CPU as such frame buffer will not work for 8514 obviously - this is exactly same approach as NEC uPD7220 and many similar graphic accelerators where frame buffer is never available in CPU address space - if you wish to set pixel then you need to program graphic accelerator.

This is beyond the scope of this topic.

On the Amiga, a fast 32-bit 68K CPU with Fast RAM can exceed the A1200's Blitter. Alice couldn't reach into the CPU's Fast RAM for memory reads. The CPU has to copy data from Fast RAM into Chip RAM.

Fast 386DX/486SX PCs can brute force their 2D games.

Originally Posted by pandy71

Nope, not me - you started this and it was way earlier than https://eab.abime.net/showthread.php...78#post1678578 .

Nope. Emiespo made a claim.

Shall we run SimCity 2000 again?

Originally Posted by pandy71

I use 8514 as example that even on PC some players realized necessity to introduce HW graphic acceleration and IBM was natural leader, foundation of standards like VGA standard for example, market closely followed IBM.
By closing 8514 specification and making it MCA attached IBM automatically killed product mass market popularity but anyway triggered some idea on PC - this idea started to grow slowly.

It's unwise to promote "hit the metal" with higher complexity than VGA on PC since hardware can change. VBE is the thin-skin SVGA access beyond VGA. Duke Nukem 3D used VBE.

You have timing-related compatibility problems with extensive Amiga chip upgrades.

PC has a better DOS Super Steet Fighter 2 Turbo port that scales with hardware performance while it's a dog on Amiga's Super Steet Fighter 2 Turbo despite using a TF1260 card.

Originally Posted by pandy71

Nope - you are constantly arguing about various VGA card performance in some various benchmarks and games like Doom - so we have endless loop of various pictures and numbers how good was ET4000 in Doom.
No Doom in Amiga, no ET4000 in Amiga, no VGA in Amiga so this is completely out of this topic.

1. My argument wouldn't change when SNES with SuperFX2 had a Doom port in 1995.

A1200 would need a fast 68030 CPU accelerator card for Doom's near full screen with low details and floor/ceiling textures.

68EC020-25 could be acceptable at a minimal cost increase and without floor/ceiling textures.

SNES has a Mode 7 Direct Color-packed pixel feature.

2. Lew Eggebrecht has low-cost DSP32 in the roadmap for low and high-end Amigas. I support the unified Lew Eggebrecht/Dave Haynie/Mike Sinz's DSP32 inclusion.

Originally Posted by pandy71

Once again - topic is about A1200 disappointment no about superiority of Tseng ET4000 over Trident 8900C - obviously you are constantly going off topic. And Trident was more popular because its lower price so obviously not everyone was interested in Doom performance.

Nowadays large part of PC market are embedded graphic controllers (IMHO they are dominating market).

That's a flawed argument when the gaming PC minority exceeds the Amiga install base.

PC IGP is improving with Intel and AMD duking it out in the handheld gaming PC. Both camps released their respective handheld gaming PC devices.

The current PC IGPs that are for sale in 2024 have exceeded PS4 and Xbox One.

https://www.techpowerup.com/321693/a...56-bit-lpddr5x
AMD "Strix Halo" APU, PC's 1st Apple M3 Max style SOC with 256-bit LPDDR5X-8533 and active large GPU. Strix Halo's large RDNA 3.5 40 CU GPU IGP has a 32 MB cache cushion.

In terms of concept, it's similar to PS5 SoC with 8 Zen 2 + 36 CU RDNA 2.
Strix Halo SoC has 16 Zen 5 and 40 CU RDNA 3.5.

Originally Posted by pandy71

Great - so i assume with your strong support we can expect new Amiga models with DSP on board - as you and Lew will convince CBM management.

Lew Eggebrecht has low-cost DSP32 in the roadmap for low and high-end Amigas. I support Lew Eggebrecht/Dave Haynie/Mike Sinz's DSP32 inclusion. The proposal is better than stock A1200/CD32.

Lew Eggebrecht is "Commodore management" and Commodore ran out of time and money.

Lew Eggebrecht should have been in the hot seat in 1988.

Originally Posted by pandy71

Well... Tseng for sure never was majority of PC market and this closing topic. Your analysis about Tseng market share are simply limited.
Tseng was not so popular as you think.

It's larger than Amiga's install base.

Originally Posted by pandy71

Once again - you are assuming wrongly that new PC base is bigger than old one when it was exactly opposite - for single sold 486 machine there was many 386 machines.... PC in those times was not replaced after year or two.

From https://www.intel.fr/content/dam/doc...ual-report.pdf
Intel reported the following
1. In 1994's fourth quarter, Pentium unit sales accounted for 23 percent of Intel's desktop processor volume.
2. Millions of Pentiums were shipped.
3. During Q4 1993 and 1994, a typical PC purchase was a computer featuring the Intel 486 chip.
4. Net 1994 revenue reached $11.5 billion.
5. Net 1993 revenue reached $8.7 billion.
6. Growing demand and production for Intel 486 resulted in a sharp decline in sales for Intel 386 from 1992 to 1993.
7. Sales of the Intel 486 family comprised the majority of Intel's revenue during 1992, 1993, and 1994.
8. Intel reached its 6 to 7 million Pentiums shipped goal during 1994. This is only 23 percent unit volume.

By the end of 1994, Intel's Pentium PC install base crushed the entire Amiga install base of 4 to 5 million units!

By the numbers, Intel's unified X86 PC platform is a monster compared to the Amiga i.e. it mirrored the USA military might against the smaller German military during WW2.

It's larger than Amiga's install base.

Originally Posted by pandy71

I can read so this is my question - provide detailed list of operations where ET4000 assist CPU in a way not present on other VGA chips.

It's SimCity 2000 time.

Originally Posted by pandy71

And CPU assistance is not HW acceleration where HW replace CPU utilization.

On the Amiga, a fast 32-bit 68K CPU can exceed the A1200's Blitter.

386DX-40 will beat 68EC020 at 14 Mhz with FastRAM into oblivion.

Originally Posted by pandy71

You are fool saying this - if HW was slow then you could use CPU - nobody force you to use slower HW. This should be clear also for you.

You're foolingly equating AGA tech with 8514.

[hammer]

Originally Posted by robinsonb5

It absolutely has to do with modern operating systems being more complex - but that runaway complexity is itself a huge problem, and one with no easily attainable solution in sight.

And yes, modern Linux is just as bad - booting it from spinning rust is very much like booting an Amiga while its boot volume is validating.

Boot with NVMe.

[oscar_ates]

68030/fast ram combo was needed to compete with the PCs, however it never came to table for A1200 at commodore times. Interestingly 68030 was on the table for the ESCOM's darth vader helm machine, when PCs were moving to Pentium-100.

[Thomas Richter]

The bare bone ET4000 (without W32) does not really have much to offer in terms of accelerator, though it has a very fast ISA bridge which gave it a market advantage. Probably the only feature it has is color expansion on write (i.e. you could write to it in "planar", and it would expand the foreground and background color for chunky displays). It also offered a linear apperture window which probably helped some applications - you would not need segment registers - but that was not special. That is about all I remember.


The accelerator came with the ET4000W32 you find on the Merlin card.

[Promilus]

@pandy - add him to ignore list, this thread looks a lot nicer when there's no wall of text. Mine, yours, his - mostly his...
@Thomas_Richter - RTG on PiStorm hardly uses ANY hardware acceleration of VideoCore (Broadcom iGPU), afaik it does use only scaling etc. Bare minimum. Because nobody did create actual videocore driver in AmigaOS space. And that was exactly the point of exposing all RPi hardware to Amiga address space.

Also regular Tseng ET4000 only does some fairly limited bit masking and logic operations on data. So there's at least something, but not much.

[sandruzzo]

I've lost the plot: where are we going with this post? It was quite interesting...

[Dunny]

Originally Posted by dreadnought

Whoa, much hyperbole

Well, maybe a little

Seriously though, are you guys sure this is only because Windows is so unmitigably awful, and has nothing to do with the small matter of modern O-systems being somewhat more complex than the +30 years old ones?

Which is of course the problem. Those old operating systems - such as AmigaOS of course - still boot in the same time they did 30 years ago because they've not been updated to keep pace not only with modern hardware (to an extent) but also with modern practices. PCs just have to do a lot more than they used to in order to ... do the same shit we used to do, but while preventing bad actors and rogue systems from getting in and wreaking havoc.

There's no wonder it's all so much slower now, even if you don't take into account the very obvious bloat in pretty much everything.

[Thorham]

Originally Posted by Dunny

There's no wonder it's all so much slower now, even if you don't take into account the very obvious bloat in pretty much everything.

On my current peecee everything is blazingly fast, in spite of current bloatware.

[pandy71]

Originally Posted by hammer

That's a false narrative.

I'll post it again:

You again posting useless random information's totally unrelated to topic.

Originally Posted by hammer

1. My argument wouldn't change

Yes, i agree - you will not stop flooding this thread with misleading, topic unrelated numbers, pictures, quotes etc.
Good luck with your bat and dead CBM corpse.

Originally Posted by hammer

ColdFire is rubbish for Amiga's legacy game use case.

This depends only on CF clock.
CF in legacy code is still comparable to 68030 speed.

Originally Posted by hammer

That's not an "out of the box" experience.

It can be made to be like this and original Amiga was also not "out of the box" experience. Poor argumentation.

Originally Posted by hammer

This topic is about games.

From now? Until you change rules next times?

Originally Posted by hammer

There's an economic of scale issue with 3rd party Amiga CPU accelerators. Each Amiga accelerator card design is unique for each Amiga model. There's no common compute module for A500/A1500/A2000/A1200 when compared to Raspberry Pi's small board computers(SBC). The low-risk additional cost is the gateway PiStorm adapter board for each Amiga model.

As always there is economic of scale - this is truism...

Originally Posted by hammer

Imagine, a common 68060 compute module that can be used from A500/A1500/A2000/A1200 and standalone Linux68K. 68060 accelerator card wouldn't be bound by A1200's smaller market size but by the combined A1200/A500's larger market size. The common 68060 compute module would be spread to other 68K based systems. Replace 68060 with another big-endian CPU family. That's a hindsight after RPi form factor innovation.

And still this is dead end as 68060 is no longer produced and it is outdated.

Originally Posted by hammer

For B2B relationships, IBM had released 8514 hardware documentation to MS. Windows 8514 driver doesn't fully rely on 8514's AI since it's insufficient.

Fine - provide this B2B documentation, TIA.

Originally Posted by hammer

You have timing-related compatibility problems with extensive Amiga chip upgrades.

Skipped all above as not topic related.

And you don't need to have any problems with timing if you do things correctly.

Originally Posted by hammer

You're foolingly equating AGA tech with 8514.

Once again - skipped irrelevant topic flood.

I never put equality between 8514 and AGA - if you get it in that way you must be fool - all i've wrote is that 8514 is first time approach to settle graphic acceleration standard in PC market.
Failed due IBM marketing idea but still it shows that HW acceleration can't be replaced by pure CPU power. Nowadays this is obvious - it was not so obvious in past and you are living proof of this.