@WolfToTheMoon

Definitely an interesting opportunity existed at the time for C= to focus on two business streams:
1. The C stream of more standardised UNIX workstations in multiple SKUs, and
2. The A stream of multimedia Amiga OS computers in multiple SKUs.

Hardware and software developments in one stream could help the other, and eventually they could/would merge the streams into a single stream of capable computers, running a single UNIX-based AmigaOS, in multiple SKUs.

_________________
All the way, with 68k

@agami

I believe Lorraine was somewhat of a knee-jerk reaction to the introduction of the Macintosh.
And C900 and Zilog business was Tramiel's idea, so it was probably cancelled after he left having lost support at the upper management level. Notice that even the plan for Zilog to manufacture some of the C= chips fell through, and that wasn't directly tied to the C900.
Dave Haynie said that the decision to cancel the C900 was not very popular inside the company as it was felt that the A1000 and C900 were not direct competitiors and were aiming at different markets - I think, it was also stupid because the C900 development was basically finished and it reuslted in all that time and money being wasted, with only the case being reused for the A2000.

One could argue that, if you cancel some of the questionable 8 bits project, possibly even C128, and concentrate on the UNIX and AmigaOS lines... you have a much better line-up than historically in the late 80s. And if Commodore buys Zilog, that gives them inhouse 16/32 bit CPUs that are cheap to make and reasonably competitive until the early 90s - so they don't have to be overly reliant on the 65xx and M68K.

_________________

@cdimauro
https://youtu.be/PbVjNInzrw8?t=1357

"Well, Apparently 50 prototypes were made of which this is one of them and there are other examples of prototypes out there which are working. So let me show you one of those examples now. This one is owned by Sunto from VintageComputer.ca and I think he's got the high resolution card here. That's what's being demonstrated and trade show demos."

He could be wrong about the high resolution card. The Demo goes on to show 1024x800 on screen in what looks like Italian. It still could be advertising the HR C900 with a LR C900 but that wouldn't be very good for marketing.

agami Quote:

Definitely an interesting opportunity existed at the time for C= to focus on two business streams: 1. The C stream of more standardised UNIX workstations in multiple SKUs, and 2. The A stream of multimedia Amiga OS computers in multiple SKUs. Hardware and software developments in one stream could help the other, and eventually they could/would merge the streams into a single stream of capable computers, running a single UNIX-based AmigaOS, in multiple SKUs.

CBM did go on to create one universal Amiga hardware that could use AmigaOS and Unix. Where does this "single Unix based AmigaOS" idea come from though? How does that work?

@matthey

Quote:

CBM did go on to create one universal Amiga hardware that could use AmigaOS and Unix. Where does this "single Unix based AmigaOS" idea come from though? How does that work?

They did release it, but sadly 6 years after the C900.

Atari also tried something similar with the TT030 and companies wanted it but same as Amiga, the support was lacking and it was too little, too late.

_________________

WolfToTheMoon Quote:

They did release it, but sadly 6 years after the C900. Atari also tried something similar with the TT030 and companies wanted it but same as Amiga, the support was lacking and it was too little, too late.

There were 2 attempts at using Unix OS on universal Amiga hardware with Unix OS.

1988-1989 Amiga 2500UX (development system?)
https://bigbookofamigahardware.com/bboah/product.aspx?id=21

1990 Amiga 3000UX
https://bigbookofamigahardware.com/bboah/product.aspx?id=26

The attempts were much later than the C900 but also more professional, other than using an aging ECS chipset. More competition no doubt reduced demand. Apple also had their Unix and was better known for their marketing, support and service which CBM lacked.

AMIX (Amiga Unix)
https://en.wikipedia.org/wiki/Amiga_Unix

A/UX (Apple Unix)
https://en.wikipedia.org/wiki/A/UX

ASV (Atari Unix)
https://en.wikipedia.org/wiki/Atari_TT030#Details

A/UX was the first to market in 1988 but was low quality until A/UX 3.0 in 1991. CBM seemed hesitant to release AMIX which could have been released in 1988 for the Amiga 2500(UX) and likely would have already been better than A/UX. The main attempt was the Amiga 3000UX in 1990. Atari's ASV was late being introduced in 1992 and support was pulled soon after. Linux was released in 1991 using cheap 80386 hardware which finally provided not only a large flat address space but also a MMU. It was still a pain to program compared to the preemptive multitasking 68008 Sinclair QL Linus Torvalds owned but the PC hardware was cheap and available. It was awhile before Linux became good but it changed Unix forever. The 68k Amiga did get an early port of Linux that was pretty good but CBM committed financial suicide while Apple left the 68k for what looked like greener pastures with PPC before succumbing to the cheap x86 PC hardware monster too.

@matthey

Quote:

matthey wrote: @cdimauro https://youtu.be/PbVjNInzrw8?t=1357 "Well, Apparently 50 prototypes were made of which this is one of them and there are other examples of prototypes out there which are working. So let me show you one of those examples now. This one is owned by Sunto from VintageComputer.ca and I think he's got the high resolution card here. That's what's being demonstrated and trade show demos." He could be wrong about the high resolution card. The Demo goes on to show 1024x800 on screen in what looks like Italian. It still could be advertising the HR C900 with a LR C900 but that wouldn't be very good for marketing.

OK, I see what do you mean now.

I believe that it's the second, because the displayed characters (especially) and graphics doesn't look in high resolution.

Yes, it's a very bad advertising because the system is clearly very slow.

@matthey

Quote:

matthey asked: Where does this "single Unix based AmigaOS" idea come from though? How does that work?

Were they to have focused on two streams (Zilog/UNIX and m68k/Amiga OS) in the late '80s, they would have essentially beat Steve Jobs to the mark of having systems for "Pros" and systems for the "Creative Consumer".

It follows that if this version of C= were successful enough to survive past 1994; to compete with Windows, BeOS, and others, Amiga OS users would demand parity in memory management, multi-user environments, eventually 64-bit support, multi-core, etc. And with UNIX workstations from Sgi, Sun, adding more AV features, it would become impractical and uneconomical to maintain two separate, yet overlapping code bases.

Coupled with the fact that many bespoke RISC micro-architectures didn't survive the x86-64 onslaught, C= workstations and Amiga dream machines would eventually end up using the same CPU architecture. In such a world, C= would be better off with an OS X like approach with a C= UNIX foundation, and an Amiga UX layer. One OS that can be both serious enough for the science lab, and friendly enough for everyday use by everyday people.

Last edited by agami on 23-Aug-2024 at 07:35 AM.

_________________
All the way, with 68k

@agami

Quote:

Were they to have focused on two streams (Zilog/UNIX and m68k/Amiga OS) in the late '80s, they would have essentially beat Steve Jobs to the mark of having systems for "Pros" and systems for the "Creative Consumer".

If C= is successfuly selling reasonably priced UNIX workstations in the latter part of the 80s, I don't think there would have been a market for NeXt. That is one interesting repercussion of the C900 line living on - Apple would face an interesting conundrum in the 90s, and in this version of history, maybe Jobs doesn't return to Apple.

Quote:

It follows that if this version of C= were successful enough to survive past 1994; to compete with Windows, BeOS, and others, Amiga OS users would demand parity in memory management, multi-user environments, eventually 64-bit support, multi-core, etc. And with UNIX workstations from Sgi, Sun, adding more AV features, it would become impractical and uneconomical to maintain two separate, yet overlapping code bases. Coupled with the fact that many bespoke RISC micro-architectures didn't survive the x86-64 onslaught, C= workstations and Amiga dream machines would eventually end up using the same CPU architecture. In such a world, C= would be better off with an OS X like approach with a C= UNIX foundation, and an Amiga UX layer. One OS that can be both serious enough for the science lab, and friendly enough for everyday use by everyday people.

yes, eventually they were to merge... IF Amiga machines are indeed launched in the "C900 survives" parallel universe. I could see an idea to launch Amiga as an game console.

_________________

cdimauro Quote:

OK, I see what do you mean now. I believe that it's the second, because the displayed characters (especially) and graphics doesn't look in high resolution. Yes, it's a very bad advertising because the system is clearly very slow.

Performance is somewhat relative. The 68000 Amiga seems slow to most Amiga users today. In 1985, the 68000 Amiga was amazing though.


agami Quote:

Were they to have focused on two streams (Zilog/UNIX and m68k/Amiga OS) in the late '80s, they would have essentially beat Steve Jobs to the mark of having systems for "Pros" and systems for the "Creative Consumer". It follows that if this version of C= were successful enough to survive past 1994; to compete with Windows, BeOS, and others, Amiga OS users would demand parity in memory management, multi-user environments, eventually 64-bit support, multi-core, etc. And with UNIX workstations from Sgi, Sun, adding more AV features, it would become impractical and uneconomical to maintain two separate, yet overlapping code bases. Coupled with the fact that many bespoke RISC micro-architectures didn't survive the x86-64 onslaught, C= workstations and Amiga dream machines would eventually end up using the same CPU architecture. In such a world, C= would be better off with an OS X like approach with a C= UNIX foundation, and an Amiga UX layer. One OS that can be both serious enough for the science lab, and friendly enough for everyday use by everyday people.

It all sounds a little too utopian to me. AmigaOS would have likely become an embedded OS which it practically is right now. There would have been the option to create a 64-bit SMP version of it but I wouldn't guarantee it would be based on Unix/Linux. Let's not forget that Unix evolved from less than where the AmigaOS started.

https://en.wikipedia.org/wiki/Unix#Overview Quote:

At first, Unix was not designed to be portable or for multi-tasking. Later, Unix gradually gained portability, multi-tasking and multi-user capabilities in a time-sharing configuration.

The problem was that the AmigaOS did not evolve or proliferate being hidden away. AmigaOS 4 created a new incompatible "desktop" AmigaOS but competitive desktop features were left out. The AmigaOS could have been replaced by something else like QNX or even Windows but probably not unless the architecture changed. I believe most Amiga fans would like to see organic development of the AmigaOS and Amiga chipset which requires organic development of the 68k as well. The Amiga1 shows what happens when moving away from the Amiga heritage. Maybe CBM could have pulled off a change in architectures and OS like Apple but they likely would have lost a lot of customers like they did from the C64 to the Amiga. Compatibility became more important and it was possible with enhancements unlike earlier hardware. It is easier to market to existing users and fans rather than trying to create a new market. Where have I heard that before? Isn't the solution still to evolve and proliferate on affordable 68k hardware? Why does this require Unix/Linux?


WolfToTheMoon Quote:

If C= is successfuly selling reasonably priced UNIX workstations in the latter part of the 80s, I don't think there would have been a market for NeXt. That is one interesting repercussion of the C900 line living on - Apple would face an interesting conundrum in the 90s, and in this version of history, maybe Jobs doesn't return to Apple.

CBM was targeting low priced UNIX servers and workstations. NeXT was higher end. I have doubts about the Z80000 even if CBM had bought Zilog and starting with the Z8001 was a poor choice. C900 hardware was based on serial port networking where NeXT used Ethernet and had TCP/IP. NeXT likely had better BSD based software which is the same base as the modern macOS uses today. The C900 needed a lot of upgrading to end the NeXT market. The Amiga3000UX was nicer Unix hardware and more competitive as other Unix competition arrived too. NeXT hardware prices came down to be reasonable. The 1990 released 68040@25MHz NeXT workstation started at $4,995 which was high end Mac territory. The Amiga 3000UX with 68030@25MHz started at $4999 and it did not include the A2410 graphics card for color or an Ethernet card. The value was ruined by needing an expensive graphics card because the Amiga chipset was not capable enough after failing to upgrade it adequately.

WolfToTheMoon Quote:

yes, eventually they were to merge... IF Amiga machines are indeed launched in the "C900 survives" parallel universe. I could see an idea to launch Amiga as an game console.

The Amiga was launched as a console called the CD32. Unix wouldn't have helped as it uses more memory and low end hardware MMU usage kills performance.

Last edited by matthey on 24-Aug-2024 at 01:33 AM.

@WolfToTheMoon

Quote:

If C= is successfuly selling reasonably priced UNIX workstations in the latter part of the 80s, I don't think there would have been a market for NeXt. That is one interesting repercussion of the C900 line living on - Apple would face an interesting conundrum in the 90s, and in this version of history, maybe Jobs doesn't return to Apple.

NeXT's "killer app" is its development toolchain and middleware which serves as the foundation for MacOS X and IOS. Steve Jobs focused on the "look and feel" of GUI design.
NeXTSTEP gained the best mainstream 3rd party commercial support when compared to other classic Unix distributions.

SteamOS 3.x is Arch Linux based, but Valve's optimized Proton/DXVK DirectX clone middleware, Steam store, and SteamOS handheld optimized UI are its "killer app". Valve did a good job on AMD-related power management for mobile Ryzen APUs.

Sony's GameOS is FreeBSD-based, but Sony's value-added middleware, development toolchain, and 1st party games are its "killer apps".

Google's ChromeOS and Andriod are Linux-based, and Google's value-added items are the middleware, services, and UI design.

Classic Unix itself is a yawn.

My current Linux distro are Arch Linux and HoloISO (SteamOS 3-Holo).

For Unix with performance, MMU's TLB cache to be large with high-performance MMU implementation.

Last edited by Hammer on 24-Aug-2024 at 09:28 AM.
Last edited by Hammer on 24-Aug-2024 at 09:26 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 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@matthey

Quote:

CBM was targeting low priced UNIX servers and workstations. NeXT was higher end.

Yes, but we're comparing apples and oranges here.... C900 is a 1985 machine. NeXT was out in 1988. By 1988, C900 would have been cost reduced and a possible C9000, with the Z80000 introduced, which could have been powerful enough and cheap enough(plus, there's 3 years of software base on the C900) to make the NeXt computer uncompetitive on the market.

Quote:

The Amiga 3000UX with 68030@25MHz started at $4999 and it did not include the A2410 graphics card for color or an Ethernet card. The value was ruined by needing an expensive graphics card because the Amiga chipset was not capable enough after failing to upgrade it adequately.

which is already solved on the C900(high res graphics card). And I feel that if Commodore is producing the Z80000 inhouse, they'll get a better price than 68030 from Motorola.

_________________

@matthey

Quote:

1988-1989 Amiga 2500UX (development system?) https://bigbookofamigahardware.com/bboah/product.aspx?id=21

It's time to end this Unix farce.


Quote:

From "Commodore - The Final Years", The Unix Group (skip) Johann George was an immaculately dressed programmer who came from the Coherent Unix team. (skip) Dave Haynie completed the A2620 accelerator card in early 1988, complete with the 68851 MMU chip, (skip) With the core of Amiga Unix complete, George would hire a team to help finish the first release of Commodore’s version of Unix 5.3 for the Amiga 2500. On April 27, 1988 he put out the call for experienced Unix programmers. (skip) Commodore's engineering group jumped from 49 engineers in 1987 to 86 engineers in 1988. It was the C900 project all over again, one that had proven costly and helped bring Commodore close to its destruction. “That was definitely a failure of management, one that I didn't fully appreciate at the time, but in retrospect I can see it was just fatal,� says Nesbitt. (skip) Amiga 3000 Stalled After missing CeBIT, it seemed like the engineers did not have the same sense of urgency they had when putting out the Amiga 500 and Amiga 2000 under former CEO and President, Thomas Rattigan. (skip) However, under the leadership of Irving Gould and Henri Rubin, there was no such urgency. Both men lacked the experience to get their soldiers moving. (skip) After Gerard Bucas and Bob Welland left at the end of February, the Amiga 3000 project fell into turmoil. (skip) One positive side effect of Haynie’s work with the A2630 that affected the A3000 design is that he became intimately familiar with the 68030 processor. Prior to this, the A3000 was to use the 32-bit 68020. But Haynie soon concluded that he preferred the faster processor, with its built in MMU. (skip) Since 1987, Luck had built up his Boing! company around selling a piece of software called the X Window System for the Amiga Unix operating system (Amix), as well as a three-button Amiga mouse. (skip) Commodore had also begun designing network adapter cards for the Amiga in mid-1988, which would be used for networking under Amix or AmigaOS. (skip) And Luck himself had negotiated a deal with Commodore whereby he would include copies of Amix with his X Window System. His X Window sales would also spur sales of the A2024 monitor. The relationship was mutually beneficial but his plans depended on Commodore delivering the products. (skip) Commodore had signed a contract with ADI in 1988 to manufacture 5000 A2024 monitors. However, the problematic design meant that production started and then stopped several times. In May 1990, ADI started again and finally achieved 5000 units later in the year. It also took Commodore until the end of the year to finally produce an updated manual for the A2024 that addressed the new AmigaOS 2.0. (skip) And by 1990, the Amiga was becoming known for its color, something the A2024 was not capable of producing. “I think it wasn't really in the mainstream of Commodore's business,� says Jeff Porter, referring to gamers. (skip) In the end, the few A2024 units that made it into customers’ hands were mainly used with A2500 machines and Unix, where a high resolution display was useful for displaying lots of windows. Amiga Networking Work Group The other part of Commodore’s Unix plan revolved around networking abilities under Amix. In mid-1988, Commodore engineer Ian Kirschmann was finishing off a network adapter for the A500 called the A560, along with a card called the A2060 for the Zorro II port. Jeff Porter had decided to embrace the ARCNET communications protocol over Ethernet, due to the lower cost of ARCNET. In fact, the entire bill of materials for the cards came to only $42, meaning they could sell the cards quite profitably in the $100 or higher range. (skip) Amix, which was built upon the Coherent operating system, had no network stack. It was up to Johann George and his group of programmers to implement one. Meanwhile, Andy Finkel and his group would implement a TCP/IP stack to use with AmigaOS, although little more than a network adapter driver came from that effort. The most coherent effort began with Dale Luck himself, who established the Amiga Networking Work Group at the June 1989 Amiga Developer’s Conference in San Francisco. His goal for the group was to make the Amiga a fully functioning partner in a network that included Macintosh, PC, and Unix machines. This included file servers, shared printers, and even an email client. (skip) The slowness of getting these products to market affected not only Dale Luck. Henri Rubin had been eager to show this impressive collection of technology to Irving Gould and the board of directors as evidence of his effectiveness. Instead, it looked like Rubin was fumbling many or most projects that he started at Commodore. (skip) A3000UX Released Commodore had planned to release the A3000UX machine in September 1990. However, the first units arrived to dealers in January 1991 and Commodore began gearing up for the launch. A few months later, in March 1991, they were available for purchase in the US. During its development, Henri Rubin and Johann George had demonstrated it at shows. Despite the impressive accomplishment of running Unix on Amiga hardware, the computer did not appeal to Commodore’s core audience, who preferred AmigaOS to run common applications and play games. (skip) The A3000UX was not well received by the press. Unix World magazine reviewed the system in depth and found that, although it had outstanding graphics and sound compared to competing Unix workstations, it was no bargain at $6998, excluding monitor and tape drive. The reviewer found fault mainly with the outdated 68030 chip (which had been superseded by the 68040) and the sparse software available. Around 100 of the most popular Unix programs had been compiled to Amix by Commodore, and most of those did not use the Amiga’s remarkable sound and graphics capabilities.

NextSTEP's third-party mainstream commercial software library from MacOS's third-party list is far superior to Coherent and AMIX. Steve Jobs is setting up a "perfect storm" with NextSTEP-based MacOS X.

Quote:

From "Commodore - The Final Years", June 1991 marked a turning point for Bryce Nesbitt’s career at Commodore. He had previously taken a stand that if Commodore failed to announce the A3640 board in a timely fashion, he would resign from the company. “We had it! It worked! It was fast. It was the latest processor and we were the first ones to have it, but the thing that was revolutionary about it was the memory management unit.� Commodore had failed to publicize the board at CES that summer and Nesbitt felt he had no choice. “That in fact precipitated my leaving Commodore,� he says. Although he did not immediately resign, he began telling coworkers his disappointment in the company after CES

Commodore management's A3640 release delay. Around March 1991, the pro-PC administration was running the Amiga R&D.

Last edited by Hammer on 24-Aug-2024 at 11:17 AM.
Last edited by Hammer on 24-Aug-2024 at 10:53 AM.
Last edited by Hammer on 24-Aug-2024 at 10:51 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 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@WolfToTheMoon

1. During AMIX's development, Coherent-based AMIX is missing a network stack from Coherent.

Commodore added a network stack for AMIX.

2. Commodore's Unix group doesn't have a revenue base to support itself.

_________________
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 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@Hammer

Quote:

@WolfToTheMoon 1. During AMIX's development, Coherent-based AMIX is missing a network stack from Coherent. Commodore added a network stack for AMIX. 2. Commodore's Unix group doesn't have a revenue base to support itself.

All of those things could have been solved by launching C900 in 1985 - it took Amiga(and Macintosh and Atari) a few years to get going, it's unreasonable to expect everything sorted out in 1985 on launch day.

_________________

@WolfToTheMoon


Quote:

All of those things could have been solved by launching C900 in 1985

1. Coherent's Mark Williams Company was closed in 1995.

2. Coherent version 4 was released in 1992, Coherent adds support for X11 and MGR windowing systems. That's late.

Dale Luck has ported X-Windows to the Coherent-based AMIX branch.

3. PC Magazine called Coherent 3.0 a "time capsule" that captured the state of Unix in the late 1970s, without support for mice, LANs or SCSI disks, good for learning basic Unix programming but not for business automation. - Christian, Kaare (11 December 1990), PC Magazine.

Coherent is a joke next to NeXTSTEP.

Steve Jobs NeXTSTEP team was doing the hard work preparing for their Apple return and MacOS X nuke against other commercial Unix workstations.

The UI from Coherent is LOL.

Last edited by Hammer on 24-Aug-2024 at 12:27 PM.
Last edited by Hammer on 24-Aug-2024 at 12:22 PM.
Last edited by Hammer on 24-Aug-2024 at 12:18 PM.
Last edited by Hammer on 24-Aug-2024 at 12:17 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 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@WolfToTheMoon

Quote:

WolfToTheMoon wrote: which is already solved on the C900(high res graphics card). And I feel that if Commodore is producing the Z80000 inhouse, they'll get a better price than 68030 from Motorola.

Amiga custom chipset's flat memory model was designed with a 68000's flat memory model, hence Z8000's 16-bit segmented memory model wasn't suitable.

https://groups.google.com/g/comp.arch/c/9Z6kS5ytRTU/m/8TrqL-hoS64J
Posted Date: December 1987.
"The Z80,000 was put on market just about 8 months ago."

Delays, it's too late and has supply problems.

_________________
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 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@Hammer

Z80000 was late too market, yes... It might have been different if Zilog was bought out.

check this out

https://www.techmonitor.ai/technology/zilog_offers_z80000_version_at_rock_bottom_26_in_large_lots

Quote:

The commercial Z80320 version will come in a 68-pin plastic leaded chip carrier and will be offered in 8MHz and 10MHz versions, with samples of a 15MHz version planned by year-end. The 8MHz version carries the $25.95 tag in large lots – the price for 100-up will be $40.48, while the 10MHz version is $33.51 for 5,000-up, $52.28 when you order 100. That compares with a very cheap $80 for 1,000 or more of the ceramic version, and $300 for a Motorola 68020, $850 for a 25MHz NatSemi NS32532.

It's much cheaper than the 68020 but faster.

_________________

@WolfToTheMoon

For Chip VRAM Super A500 for 1988, Commodore's 68020-16 price is $77

From Commodore - The Final Years.
Quote:

System Plans for 1988 Porter dubbed the new computer the Super A500. (skip) Commodore would have to spend $77 for the 68020 chip plus $100 for VRAM on top of the basic $230 cost of the original A500 (minus redundant costs for memory and processor). (skip) After Dave Haynie completed the A2000-CR project, it was time for him to get to work on the next iteration of Commodore’s high-end line of Amigas. The new project, which the engineers dubbed the A3000, would also reside in the new A2001 case. And much like the Super A500, it would use the hi-res Amiga chipset, a Motorola 68020 chip running at 14.2 MHz (double the clock speed of the A2000), high density floppy disks storing up to 1760 KB, 2 MB of VRAM, and a SCSI interface.

2MB Chip VRAM is crazy expensive let alone 1 MB Chip VRAM. PC world wasn't VRAM crazy.

VRAM config was dumped.

Later, Hombre's Agnus replacement could access both 32-bit/64-bit VRAM and 32-bit DRAM in a hybrid design.

3DO's MADAM (Agnus role) could access both 32-bit VRAM and 32-bit DRAM in a hybrid design.

Both 3DO and Hombre have lessons from "Super A500".

A hybrid design could enable Agnus/Alice replacement the ability to access small storage/high bandwidth VRAM and larger storage/lower bandwidth DRAM.

The valid solution is between DRAM vs VRAM extremes i.e. a hybrid memory storage model.

In modern GPUs, you could have high bandwidth 64 MB L3 cache and slower bandwidth GDDR6-20000. Without delta color compression, 900p/1080p framebuffers fit within 32 MB ESRAM for Xbox One.

Xbox Series X has 10 GB in 320-bit bandwidth and 6 GB in 192-bit bandwidth in a hybrid UMA.

Techmonitor doesn't have access to Commodore's discount prices with Motorola.

68EC040-25 could be pushed into near A500's $699 price, but the problem is missing the support chips. Commodore management made sure no mid-price Amiga would embarrass Commodore's i386DX-33/Am386DX-40 PCs.

The price for 68EC040-25 is similar to i386DX-33 and Am386DX-40. Motorola tried to support Commodore, but Motorola is a bit careless with 68EC040's cache handling and Commodore management is pro-PC.

Imagine, i386DX-33/ Am386DX-40 mid-priced range Amiga AA with 68EC040-25. That's like 386DX-40 priced PC with 486SX-25 level compute power which is pretty good.

68LC040-25 and 486SX-25 have similar pricing.

Amiga AA with 68EC040-25 + support chips could easily evolve into a 68EC060-50 with a similar price in 1994. 68EC060-50 for $100 in 1994 is not bad i.e. near Pentium-like power (minus FPU) at 1992's i386DX-25/Am386DX-40 prices.

Dave Haynie argues his new support chips can handle lower cost 68EC040-25's cache issue.

Last edited by Hammer on 24-Aug-2024 at 04:26 PM.
Last edited by Hammer on 24-Aug-2024 at 04:23 PM.
Last edited by Hammer on 24-Aug-2024 at 04:16 PM.
Last edited by Hammer on 24-Aug-2024 at 04:12 PM.
Last edited by Hammer on 24-Aug-2024 at 04:08 PM.
Last edited by Hammer on 24-Aug-2024 at 04:06 PM.
Last edited by Hammer on 24-Aug-2024 at 04:01 PM.
Last edited by Hammer on 24-Aug-2024 at 03:58 PM.
Last edited by Hammer on 24-Aug-2024 at 03:53 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 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@Hammer

Quote:

Commodore's 68020-16 price is $77

That is still 2-3 times more than the quoted price for the Z80000, and it lacks the Zilog's built in MMU, so you'd have to add another chip to match it. 68030 was the better match.

_________________

@WolfToTheMoon

Z80000's mass production was too late when the A500 was released in 1987.

Without an extensive software library, Z80000 would be treated like any new CPU ISA offering.

MMU was not a major issue when Commodore created its custom MMU for 68020 before Motorola released the 68851 MMU.

The original A2620 card includes a Commodore's custom MMU instead of Motorola's late 68851 MMU. This caused A2620 to be redesigned, hence wasted R&D resources.

A major factor for MIPS' success was due to high-profile design wins from the likes of SGI and DEC. Sony's PSX followed SGI's example.

In 1986, Motorola 68020 and Intel 386 had production supply superiority over Zilog's Z80000.

Macintosh II was shipped in 1987 with a 68020 CPU while Zilog solved its production problems with Z80000. You're not factoring in the system integration time.

An Apple exclusive has no problem publicly bashing Bill Mensch on 65816's supply problems.

The main reason for Apple inviting Motorola into the PowerPC camp is Motorola's experience with mass production.

AMD has used Motorola's silicon fabrication tech during the K7 and K8 eras.

Motorola is just mismanaged and lacks certain high-clock speed logic circuit knowledge.

If Motorola had some of DEC Alpha's high clock speed logic circuit knowledge, the situation would be different. There's a reason for AMD and Intel targeted DEC Alpha engineers.

IBM would later master high-clock speed designs with POWER 6 and would combine with strong IPC with POWER9. IBM POWER9 CPU design is competitive, but the platform entry price is problematic.

Leadership quality is an important factor in a company's resource governance.

Last edited by Hammer on 25-Aug-2024 at 02:52 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 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB