@ppcamiga1

Quote:

ppcamiga1 wrote: @agami The some shit again. The some scumbags instead of hard working on aros waste time on attacks on ppc, ppc users and companies thats sells ppc hardware. 24 years wasted for nothing . Want cheap hardware start working on aros. Made it as good as at least Windows XP.

Non-standard FPU with a certain PPC design was a stupid waste of R&D resources. You're the scumbag.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@matthey

Quote:

Commodore did not upgrade the Amiga hardware! They were still using 5000nm NMOS chips in 1994! The 5000nm NMOS process was from the mid 1970s so about 20 years out of date. The Mirari PPC SoC silicon is only about 10 years old and the X5000 and A1222 PPC SoCs only about 15 years old yet they are getting destroyed by ARM SBCs that cost less than the PPC SoCs. That is what old silicon does yet you are hear reading about this "acceptably" old PPC silicon?

CSG invested in 2000nm (2 micron) fabs sometime after 1986 for the C65 chipset, but it wasn't applied for the Amiga group (Commodore-Amiga Inc.)'s chipset designs. Henri Rubin's administration has the "read my lips, no new chips" directive for the 32-bit A3000 project, and bet the farm on ECS's Productivity Mode at the expense of true 256 color gaming modes.

Commodore's system engineering group (the group that designs many failed CBM's 8-bit office microcomputers, and 16-bit Z8000 C900) took over Commodore-Amiga Inc. (formerly known as Amiga Corporation) from the original Los Gatos Amiga team.

CSG's LSI team designed the VIC-20 and C64, which sustained Commodore until the Amiga switch-over. For the C65 project, the LSI team couldn't wait for Henri Rubin's rubbish leadership direction.

2-micron process node would be around MIPS 2000 and 68020 level.

Remember, the C65 chipset was a rogue unofficial R&D from CSG's LSI team.

With C65's 8 bitplane R&D experience, CSG's LSI team has spec'ed and designed AA Lisa with AAA's 16.7 million color palette.

For the 1980s and early 1990s, the PC world and Nintendo weren't crazy enough to push a 16.7 million color palette standard with many lower-cost mainstream SVGA chipsets e.g. VGA's 18-bit color palette standard is enough, while SNES has 15 bit color palette.

In 1989, AA R&D was allowed to start due to the CSG-LSI group's C65's proper 256 color display chipset surprise, which had one-upped Henri Rubin's Amiga group.

Blame Henri Rubin for the debacle. Bill Sydnes and Jeff Franks weren't any better since they both bet the farm on ECS e.g. A300/A600, and "A1000Jr," i.e. cost-reduced ECS A3000 with A2200/A2400/A3200/A3400 models.

Mehdi Ali ordered for AGA AGA-equipped A500 class machine in February 1992 that overrides Bill Sydnes' and Jeff Franks' ECS Amiga plans.

Henri Rubin departed from ECS's original intention to quickly upgrade 16-bit Amiga OCS models with productivity mode.

Henri Rubin's "betting the farm" on ECS wasted the A3000's 32-bit bus architecture.

Neither Henri Rubin nor Bill Sydnes don't understands graphics chipset lead time R&D since both lack experience in microprocessor R&D designs.

NVIDIA was co-founded by an ASIC designer for the SUN GX 3D accelerator and a former AMD microprocessor designer with a background in the LSI Coreware business (i.e. PS1!!!!).

As a graphics chip business, Commodore is not the NVIDIA gold standard.

Quote:

Ron Nicholson gave some insight on that conversation since he also worked for Apple. Apple developers were working on their own Amiga like chipset that they were trying to simplify into fewer chips but they failed. Steve Jobs thought Apple would be more successful than the Amiga developers at integrating a chipset when he made the "too much hardware" statement but he was wrong.

Steve Jobs focused on "next-gen" business software partnerships and high resolution for GUI ASAP.

Bill Gates recycled Steve Jobs' Mac GUI "next gen" plan for Windows/Excel/Word combo and crushed DOS MDA text UI establishments such as Lotus 123, WordPerfect, and Word Star.

Apple started Mac II's color display R&D (e.g., following IBM PGC's use case template, 640x480p with 256 colors) in 1985.

Mac II (color Quick Draw) and Windows 2.x retargetable graphics capabilities for their respective platforms in 1987.

Mac II's color display capability sets the groundwork for Apple's best-selling LC series from 1990 to 1993. The Amiga ECS project missed the narrow window of opportunity. i.e. A3000 is DOA in 1990.

Last edited by Hammer on 17-May-2025 at 12:10 AM.
Last edited by Hammer on 16-May-2025 at 03:03 PM.
Last edited by Hammer on 16-May-2025 at 04:02 AM.
Last edited by Hammer on 16-May-2025 at 03:48 AM.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@ppcamiga1

Quote:

ppcamiga1 wrote: 24 years wasted for nothing .

AmigaOS 4, in a nutshell.

_________________
All the way, with 68k

@matthey

Quote:

matthey wrote: Commodore did not upgrade the Amiga hardware! They were still using 5000nm NMOS chips in 1994! The 5000nm NMOS process was from the mid 1970s so about 20 years out of date. The Mirari PPC SoC silicon is only about 10 years old and the X5000 and A1222 PPC SoCs only about 15 years old yet they are getting destroyed by ARM SBCs that cost less than the PPC SoCs. That is what old silicon does yet you are hear reading about this "acceptably" old PPC silicon?

So close, still doesn't get the memo

C= did not upgrade MOS not because of bad management as MOS could have used an upgrade even back when Tramiel was still control.
They did not upgrade it because they couldn't afford even once let alone every 2-5 years needed to stay relevant.

Sure they sold C64 and A500 in volumes, but those didn't have the margins for such an investment.
The did sale Ax000 and before that PETs at high margins but never in real numbers.

Outsourcing the chip production was also no option for those low margin bedroom computers. One of the reasons why the A1200 came so late and with such bad value for money was that they had to go to HP to make AGA work.

There is also the design of it, stealing cycles from a CPU already starved for bandwidth (everything beyond 7MHz 68000) -> bad idea.
Using bitplanes on systems with 2MB or more running 8bit screens -> just plain stupid.

But, they had to do it that way as far to much SW relied on such setups and OCS was far to complicated to leave as legacy block in a modern chipset (the way VGA supports EGA/CGA modes) even if they could have made such a "modern" chipset at all (see above).

I'd wager if Amstrad had done a 68020 CPC successor in the same way they did the CPC (of the shelf chips around an of the shelf CPU) it would have been faster/better in almost all use cases (the HW, what kind of SW it would have run is another question) while still being cheaper to make.

Hence, the 68000+OCS combo was good for the A1000 in 85/86 and it was good for the A500 in 87-91 (at low and lower prices) but it also set things in motion that lead to the Amigas demise only a few years later.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Hammer

The some shit again.
The some scumbags instead of hard working on aros waste time on attacks on ppc,
ppc users and companies thats sells ppc hardware.
24 years wasted for nothing .
Want cheap hardware start working on aros.
Made it as good as at least Windows XP.

@ppcamiga1

All that ppcamiga1 want is a better AROS, and he is misunderstood. Poor thing, he doesn't even care with ppc, he just want MUI running on AROS. Please, make his dream come true!

_________________
Indigo 3D Lounge, my second home.
The Illusion of Choice | Am*ga

@thread

Mirari website is back

#6

_________________
This posting, in its entirety, represents solely the perspective of the author.
*Secrecy has served us so well*

@number6

Thanks for keeping us up-to-date.

_________________

@ppcamiga1

Quote:

ppcamiga1 wrote: @Hammer The some shit again. The some scumbags instead of hard working on aros waste time on attacks on ppc, ppc users and companies thats sells ppc hardware. 24 years wasted for nothing . Want cheap hardware start working on aros. Made it as good as at least Windows XP.

You're the real shit.

PS; My laser printer has PPC. LOL.

I would rather use MorphOS ahead of AROS since MorphOS can run AmigaOS-friendly 68K apps on PPC Macs.

Last edited by Hammer on 16-May-2025 at 03:55 PM.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

Quote:

Kronos wrote: C= did not upgrade MOS not because of bad management as MOS could have used an upgrade even back when Tramiel was still control. They did not upgrade it because they couldn't afford even once let alone every 2-5 years needed to stay relevant.

Well maybe they should have asked a bit more money for the C64 or they should have made a more professional C64 melted together with the declining PET line earlier.

C= had a lot of projects running before they bought Amiga, maybe they developed more than Apple did.

But you are right, there might be a reason for missing updates on the 8-Bit chips, they would have made great embedded MCUs if they where available as SoCs.

I was told about Irving Gould drawing out a lot of money of that company, I don't think he cared much about future development.

@OneTimer1

Well maybe, or maybe not...

Plenty of "semi professional" 8Bits existed from a wide range of companies.

Apple is pretty much the only one still standing (as a independent computer manufacturer).

C= mantra was "for the masses not the classes" and for 15 years they dominated with that approach.

Sure Gould did pay himself handsomely but in the big picture that is just another peanut on the pile created by every subdivision, every engineer and every middle management moron running their own little pet projects.

Some of them did even turn out a success, but they also shifted focus and public attention from what should have been a clear "Amiga 1st" agenda from 85 onwards.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@number6

_________________

@thread

The "first rebirth" link on the Mirari website gets constant updates with explanations of process.

So be sure to keep up with current affairs there.

#6

_________________
This posting, in its entirety, represents solely the perspective of the author.
*Secrecy has served us so well*

@agami

Quote:

years wasted for nothing

Cynical mode ON
You could say the same about anything Amiga related since at least 1994...
Cynical mode OFF


That is why I'm always wearing my rose-tinted glasses and enjoy every little bit of progress of our beloved platform. New hardware is a time for celebration, not snipe remarks.

Hammer Quote:

CSG invested in 2000nm (2 micron) fabs sometime after 1986 for the C65 chipset, but it wasn't applied for the Amiga group (Commodore-Amiga Inc.)'s chipset designs. Henri Rubin's administration has the "read my lips, no new chips" directive for the 32-bit A3000 project, and bet the farm on ECS's Productivity Mode at the expense of true 256 color gaming modes. Commodore's system engineering group (the group that designs many failed CBM's 8-bit office microcomputers, and 16-bit Z8000 C900) took over Commodore-Amiga Inc. (formerly known as Amiga Corporation) from the original Los Gatos Amiga team. CSG's LSI team designed the VIC-20 and C64, which sustained Commodore until the Amiga switch-over. For the C65 project, the LSI team couldn't wait for Henri Rubin's rubbish leadership direction. 2-micron process node would be around MIPS 2000 and 68020 level.

year | CPU | bits | transistors | process size | process type
1974 8080 8 4,500 6,000nm NMOS
1974 6800 8 4,100 ? NMOS
1975 6502 8 3,500 8,000nm NMOS
1976 8085 8 6,500 3,000nm NMOS
1976 Z80 8 8,500 4,000nm NMOS
1977 Bellmac-8 8 7,000+ 5,000nm CMOS
1978 8086 16 29,000 3,200nm NMOS
1979 8088 8/16 29,000 3,000nm NMOS
1979 Z8000 16 17,500 ? ?
1979 68000 16/32 68,000 3,500nm NMOS
1980 Bellmac-32 32 150,000 3,500nm CMOS
1984 68020 32 200,000 3,000nm CMOS
1987 68030 32 273,000 2,000nm CMOS
1990 68040 32 1,170,000 800nm CMOS
1994 68060 32 2,530,000 500nm CMOS

year | chipset | process size | process type
1985 AmigaOCS 4000nm NMOS (edit: 4000nm from https://youtu.be/WIoLuLXjK6Y?t=4189)
1992 AmigaLisa 1500nm CMOS (AGA used original NMOS Paula@4000nm)
1994 SonyPS1 500nm CMOS

The Commodore AGA chipset was limited by 4000nm NMOS to ~14MHz when Sony was using 500nm for the PS1 chipset running at 53MHz. Sony was able to put about ~1 million transistors on a chip so a 500nm process would have allowed the 3 chip Amiga chipset to fit in one chip saving cost, board space, power and transistors while increasing performance. The higher clock speed would have allowed memory bandwidth to be quadrupled from AGA. The PS1 also used VRAM which Jay Miner wanted for the Ranger chipset to about double memory bandwidth again. AGA would not have been slow if it had 8 times the memory bandwidth. Commodore likely would have survived if they saved $100 on every AA+ Amiga computer with a single chip Amiga according to Commodore documentation.

Hammer Quote:

Steve Jobs focused on "next-gen" business software partnerships and high resolution for GUI ASAP. Bill Gates recycled Steve Jobs' Mac GUI "next gen" plan for Windows/Excel/Word combo and crushed DOS MDA text UI establishments such as Lotus 123, WordPerfect, and Word Star. Apple started Mac II's color display R&D (e.g., following IBM PGC's use case template, 640x480p with 256 colors) in 1985. Mac II (color Quick Draw) and Windows 2.x retargetable graphics capabilities for their respective platforms in 1987. Mac II's color display capability sets the groundwork for Apple's best-selling LC series from 1990 to 1993. The Amiga ECS project missed the narrow window of opportunity. i.e. A3000 is DOA in 1990.

Ron Nicholson talks about this too.

https://youtu.be/gG_SvjO8w2o?t=371 Ron Nicholson Quote:

One of the reasons I went to Amiga, one of the reasons Jay Miner talked me into going to Amiga was that Steve Jobs at Apple thought doing a color computer was stupid. Atari thought that Jay Minor doing a 16-bit game computer was stupid and they thought that that was a silly idea not worth investing in. Dave Morse was willing to take the risk, raise capital and start a company to do something that some of the smartest people in Silicon Valley thought was a crazy idea and it turned out the crazy idea I don't know if it luck or whatever but you are all here playing with it.

Jay Miner wanted to expand the colors, increase the resolution, increase the memory bandwidth using dual ported VRAM and move to a 32-bit 68020 with the Ranger chipset but Commodore thought the Amiga already had too many colors and ECS productivity mode was enough. The story of the Amiga post Amiga Corporation is about ignorant blind people sabotaging visionaries.

Kronos Quote:

So close, still doesn't get the memo C= did not upgrade MOS not because of bad management as MOS could have used an upgrade even back when Tramiel was still control. They did not upgrade it because they couldn't afford even once let alone every 2-5 years needed to stay relevant. Sure they sold C64 and A500 in volumes, but those didn't have the margins for such an investment. The did sale Ax000 and before that PETs at high margins but never in real numbers.

Low end computers in general are low margin high volume. Commodore was the volume market leader with the C64. Sure, margins dropped off due to PC sales but low end PC margins were often lower perhaps even dipping into negative territory in the early 1990s. Dave Haynie jokes about Commodore paying more to make PCs than it would cost to buy them toward the end. Commodore would have been better off licensing the Commodore brand to a larger PC producer and focusing on the higher margin Amiga but they did the opposite right before the PC market collapsed.

Kronos Quote:

Outsourcing the chip production was also no option for those low margin bedroom computers. One of the reasons why the A1200 came so late and with such bad value for money was that they had to go to HP to make AGA work.

I strongly disagree. Commodore should have considered selling their fab business before it became outdated and a liability. A fabless semi business was already possible back then as demonstrated by HP producing the Lisa chip using a better chip process than CSG. Integrating the Amiga chipset and later CPU into a SoC could have improved the value of fabless semi production. According to Commodore inside documentation, a 68k Amiga SoC would have saved $100 per Amiga computer and it was not possible for CSG to produce this chip as it required at least the chip process of the AGA Lisa chip.

Kronos Quote:

There is also the design of it, stealing cycles from a CPU already starved for bandwidth (everything beyond 7MHz 68000) -> bad idea.

Memory bandwidth back then was limiting for everyone. Commodore was slow to upgrade the bandwidth because they liked cheap memory and the old chip silicon limited the chipset clock speed. The "bad idea" of sharing the bandwidth between the CPU and chipset/video was borrowed from the losers over at Apple.

https://youtu.be/WIoLuLXjK6Y?t=737 Ron Nicholson Quote:

So I learned a lot about Woz designs and one clever thing that Woz did in the Apple II that was not available in other personal computers at the time is he used something called a time division multiplex interleave memory. Instead of having a separate memory for video and a separate memory for the computer he essentially made a memory that was twice as fast as was needed for either and shared them between both. So during the video, the Apple II memory ping ponged back and forth between processor access, the 6502, and video refresh. That technology I learned about and Burrell Smith also adopted for the Macintosh and it turned out to be useful, something very similar, for the Amiga design.

The 68000 code density saved memory bandwidth over the 6502 and even more compared to fat RISC architectures like PPC. Commodore used the technology to make low cost hardware for the masses rather than higher end hardware like Jay Miner pushed for with the Ranger chipset but was denied.

Kronos Quote:

Using bitplanes on systems with 2MB or more running 8bit screens -> just plain stupid. But, they had to do it that way as far to much SW relied on such setups and OCS was far to complicated to leave as legacy block in a modern chipset (the way VGA supports EGA/CGA modes) even if they could have made such a "modern" chipset at all (see above).

Commodore internal documents from 1985/1986 warned that chunky was more efficient with Jay Miner agreeing and suggesting moving quickly to add the technology. The Amiga AA+ chipset which was the replacement for the delayed and then rushed AGA chipset would have added it. Some sources claim AA+ would have been 2 chips and 200,000 transistors which would be cheap legacy support considering this chipset has chunky and planar support for compatibility.

https://en.wikipedia.org/wiki/AA+_Chipset#Specifications Quote:

Specifications o Two Chips with 100k Transistor each. o Synchronous to video clock. o 160 - 280 pin packages. o 32 bit DRAM 60 ns Page Mode Chip Memory. o 57 MHz pixel clock. o 256 colors planar mode with AGA registers compatibility. o 4 MB 4 Mbit/s Floppy Controller with Hardware CRC floppy drives using standard technology. o Support for ALL 32 bit 680x0 CPUs. o 8x memory bandwidth increase over ECS. o 2x Blitter Performance (gets twice as many clocks as on AGA). o Rock steady 800 x 600 x 8-bit Non-Interlace 72 Hz refresh rate, Larger screens at lower refresh rates. o packed (Chunky) 16-bit color mode o FIFO serial ports with large buffer. o Increased chip ram limit up to 8 MB.

Note the 57MHz chipset clock compared to PS1 53MHz chipset clock and 8 times memory bandwidth over ECS. The CPU would preferably be clocked at 57MHz with the chipset. This is not a radically new design like AAA but a moderately enhanced AGA upgrade using better silicon for the whole chipset. It was still not using modern silicon as the chipset would easily fit in a single chip and a 68k Amiga SoC was likely already possible in 1992.

Kronos Quote:

I'd wager if Amstrad had done a 68020 CPC successor in the same way they did the CPC (of the shelf chips around an of the shelf CPU) it would have been faster/better in almost all use cases (the HW, what kind of SW it would have run is another question) while still being cheaper to make. Hence, the 68000+OCS combo was good for the A1000 in 85/86 and it was good for the A500 in 87-91 (at low and lower prices) but it also set things in motion that lead to the Amigas demise only a few years later.

Commodity parts should have 2nd source suppliers and the lack of integration results in larger boards with less performance. Room sized time share minicomputers used discreet logic chips for CPUs and were replaced by MPUs creating the personal workstation market which the 68000 MPU created. The 68000 Amiga integration brought the personal workstation down to the personal computer courtesy of Amiga Corporation. Computers are all about integration and economies of scale. Commodore bought the Amiga and after a few minor improvements stopped integrating, enhancing and innovating. End of Commodore, end of story.

Last edited by matthey on 17-May-2025 at 04:32 PM.

#6 Quote:

Mirari website is back

Was Trevor actually involved in the Mirari project or is the group sucking up to him to try to earn his favor for AmigaOS 4 support?

https://mirari.vitasys.nl/ Quote:

The Next-Gen Mainboard Designed with AmigaOS4 and MorphOS in mind ... Designed for AmigaOS4.x o Designed with AmigaOS4.x in mind o MorphOS Compatible o Runs PPC Linux just fine 🙂 o U-boot 2024 enabled o Ready for AmigaOS4.1 Update 3 o Latest hardware like NVMe and USB3*

https://mirari.vitasys.nl/our-story/ Quote:

First things first… Let’s take a moment to appreciate Trevor Dickinson and his incredible contributions to the Amiga world. His dedication and support have been instrumental in bringing us the next generation of Amiga computers, the ExecSG Kernel and the ongoing development of the Radeon Graphics drivers. Without having machines like the X1000, X5000 and A1222 this project would have never been born. Thank you Trevor, for keeping the Amiga spirit alive!

https://mirari.vitasys.nl/about-us/ Quote:

Our Team ... Trevor: Trevor has been a vocal advocate for the Amiga, raising awareness of its unique strengths and inspiring others to join the community. His financial backing has enabled the development of essential tools, libraries, soft and hardware for the next gen Amiga, ensuring its continued relevance in the modern computing landscape.

At least the Mirari team knows who is in charge of providing AmigaOS 4 support. I am not sure any amount of begging or praise for Trevor is going to gain AmigaOS 4 support until his inventory of A1222 and X5000 systems is sold. Nice try though. The problem may bet that the Mirari team is too good with their A1222 killer.

@matthey

You are hinting at the Osbourne Effect here?

#6

_________________
This posting, in its entirety, represents solely the perspective of the author.
*Secrecy has served us so well*

#6 Quote:

You are hinting at the Osbourne Effect here?

From my understanding of the Osbourne Effect, it would not apply here with two unrelated businesses/entities.

https://en.wikipedia.org/wiki/Osborne_effect

If the PPC AmigaOS 4 hardware market was a free market, it would just be competition. The announcement of new, better and lower cost products than that of a competitor is likely to have a suppressive effect on the competitor's sales but the Osborne Effect is a business shooting itself in the foot by announcing new, better and/or lower cost hardware that causes current sales to drop as people wait for the new product. Jeff Porter said Medhi Ali did it twice contributing to the demise of Commodore. The most painful time was announcing the Amiga 1200 while there was a large inventory of unsold Amiga 600s. A few Amiga 1200s made it to market before Christmas but the Amiga 600s did not sell well resulting in a catastrophic Christmas season for Commodore.

@matthey

Perhaps you already knew?

But I learned something new...4 team members plus Trevor.

Until now the vocal presence has been just Dave and Harald, so this is interesting to me.

#6

_________________
This posting, in its entirety, represents solely the perspective of the author.
*Secrecy has served us so well*

@matthey

Quote:

The Commodore AGA chipset was limited by 5000nm NMOS to ~14MHz when Sony was using 500nm for the PS1 chipset running at 53MHz. Sony was able to put about ~1 million transistors on a chip so a 500nm process would have allowed the 3 chip Amiga chipset to fit in one chip saving cost, board space, power and transistors while increasing performance. The higher clock speed would have allowed memory bandwidth to be quadrupled from AGA. The PS1 also used VRAM which Jay Miner wanted for the Ranger chipset to about double memory bandwidth again. AGA would not have been slow if it had 8 times the memory bandwidth. Commodore likely would have survived if they saved $100 on every AA+ Amiga computer with a single chip Amiga according to Commodore documentation.

Key components of the PS1's processing power, were manufactured at Oita TS Semiconductor (OTSS) and later at SCEI's semiconductor fabrication facility.

PS1's CPU was designed by LSI Logic Corp (USA).
PS1's GPU was designed by Toshiba (JP).

Sony is the system integrator, middleware/devkits, buying/governing 1st party game studios, and PR with 3rd party developers and end users. Ken Kutaragi knows his limits (designed SNES's audio DSP) and he has provided good leadership for PS1.

For economic reasons (shortage of parts), PS1 switched from VRAM to SGRAM.

Your statement about Ranger shows you didn't read the Commodore: The Final Years book. Wiki entry is mixing subject matter.

Amiga Ranger R&D was cancelled in mid-1986, and Herni Rubin ordered the original Amiga team to design the monochrome high-resolution ECS Denise. It's Henri Rubin's #metoo directive. This monochrome high-resolution ECS Denise evolved into the CSG-LSI team's four-color ECS Denise. Remember, CSG-LSI team is studying the Amiga's bitplane architecture for C65's 8 bitplane graphics chipset design. C65 started from 64 color 6 bitplanes and ended with 256 color 8 bitplanes. "Brain drain" is real when Commodore management removed the original Amiga graphics engineers.

Commodore kept Paula engineer, but Mary chip is dependent on AAA's Andrea (Agnus role chip). Andrea's key engineer also designed C65's chipset and fixed 2MB ECS Agnus bugs. There are not enough key engineers when there's an employee bottleneck. Commodore HR is flawed.

Unlike AGA's Lisa productivity mode, ECS Denise's 64 color palette productivity mode wasn't shared with OCS's 4096 color palette i.e. a quick hack that mirrors high resolution mode with C128.

From the Amiga's system engineering group, the VRAM-equipped "Super Amiga" was investigated, but 1MB to 2MB Chip RAM requirement merged with VRAM, hence making the potential "Super Amiga" expensive. This is before the Amiga system engineering group forms its own VLSI team (pinching key staff from the CSG-LSI group).

3DO MADAM (Agnus role chip) has unified access with FPM DRAM and VRAM hybrid, which is the lesson from the "Super Amiga" mistake.

Using VRAM as Chip RAM for low-performance workloads is wasteful.

AAA has additional complexity due to multiple memory configurations e.g.
32bit FPM DRAM,
64bit FPM DRAM,
32bit VRAM,
64bit VRAM.

3DO only has a single memory configuration.

The Commodore Amiga group didn't properly merge with CSG, hence failed to develop the "Amiga 1st" business plan. Blame the Commodore's executive board for the flawed business structure.

Read the "Commodore - The Final Years" book.

Quote:

Ron Nicholson talks about this too.

Steve Jobs didn't control the 1985-era Mac color R&D i.e. Steve Jobs wasn't Apple's CEO.

Without Steve Jobs, Apple couldn't properly replace classic MacOS with "next gen" MacOS X, which is based on the NextStep foundation. MacOS X countered Microsoft's "next gen" Windows NT-based Windows XP. Both NextStep and Windows NT have their origins in the late 1980s.

NVIDIA's co-founders have a technical proven record with AMD's microprocessor R&D, LSI Coreware (MIPS) management, and SUN's GX 3D custom ASIC accelerator. NVIDIA has the technical skills at the exclusive board level to compete against SGI (MIPS, IRIS GL, OpenGL).

Last edited by Hammer on 23-May-2025 at 01:15 AM.
Last edited by Hammer on 17-May-2025 at 02:05 AM.
Last edited by Hammer on 17-May-2025 at 02:04 AM.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB