Commodore has given us some pretty questionable innovations as well as projects that never came to life that should have brought the Amiga platform up to speed. It is time to see how the platform could have evolved, taking into account the real needs (of programmers, first and foremost) and in light of the technology available at the time.

English: https://www.appuntidigitali.it/23323/missed-opportunities-to-improve-the-amiga-chipset-6-the-alternative-of-16-bit-innovations/

Commodore ci ha consegnato delle innovazioni abbastanza discutibili come pure progetti mai nati che avrebbero dovuto metter al passo la piattaforma Amiga. E' tempo di vedere in che modo si sarebbe potuta evolvere la piattaforma tenendo conto delle reali esigenze (dei programmatori, in primis) e alla luce della tecnologia disponibile all'epoca.

Italian: https://www.appuntidigitali.it/23173/le-occasioni-mancate-per-migliorare-il-chipset-dellamiga-6-lalternativa-delle-innovazioni-a-16-bit/

@cdimauro



That 1975 8088 shown on the chart must be rare and valuable 4 years before introduction.

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

Note: CMOS uses more transistors and area than NMOS but power draw is a fraction of NMOS.

It is difficult to find reliable information and I am unsure of some of the info. The Amiga chipset 5,000nm chip process size is about mid-70s from the data above and Wikichip.org says so too.

https://en.wikichip.org/wiki/5_µm_lithography_process

It wasn't good starting so far behind the tech curve but aiming higher increased costs and decreased the number of businesses that could have brought the Amiga chipset to market. If Amiga Corporation had been able to obtain financing and stay independent, the low aim likely would have looked better. Development was so much more difficult and expensive back then that a small fabless semi business was too far ahead of its time. Aiming a little more modern likely would have allowed a higher transistor budget, higher chipset clock speeds and other goodies from inception. It's hard to believe the 68k Amiga was evolutionary using decade old chip technology. The original 68000 chip process was almost as outdated by the mid-1980s too. The 68000 was evolutionary itself creating the affordable personal MPU workstation market by antiquating timeshare minicomputers just half a decade earlier. That time was almost up so it was relegated to king of the 16/32 bit embedded market for another decade or more and provided a little desktop service too.

Last edited by matthey on 02-Sep-2024 at 02:06 AM.
Last edited by matthey on 02-Sep-2024 at 02:01 AM.
Last edited by matthey on 02-Sep-2024 at 02:00 AM.

@cdimauro

Under Henri Rubin's administration,
1. Rubin focused on add-on improvements for the Amiga OCS e.g. Amber, A2024 high res monitor, X86 bridgeboards, Unix and R&D.

Henri Rubin didn't focus Amiga's core graphics upgrades.

Lower ranks such as Jeff Porter was pushing for 8 bitplanes with 16 million colors and Dale Luck was pushing for 16-bit (65,536) color.

3DO's sustained performance mode is a palettized 16-bit (65,536) color mode with 16.7 million colors palette.

Jeff Porter's "8 bit planes with 16 million colors" push has materialized as AA or AGA.

Henri Rubin was moved to Commodore board and CDTV group before exiting from Commodore. Henri Rubin has no experience with computers.

AA600 (A1200) was spec'ed by Jeff Franks from Commodore PC clone.

AA600 ordered by Mehdi Ali in Feb 1992 against Bill Sydnes direction.

Commodore Germany wanted the AA500 with a hard disk ASAP.

A300(A600)'s PCMCIA is just Commodore's management's anti-GVP.


2. As a weak #metoo move, Henri Rubin ordered "monochrome hi-res Denise" and canceled Amiga Ranger R&D in mid-1986.



3. VGA's release around July 1987 led to "monochrome hi-res Denise" R&D wastage and moved to color high res Denise.

Commodore LSI group rejected 8 color registers with shared 4096 color palette color high res Denise. Commodore LSI group's excuse is "too complex".

Commodore LSI group compromised with 4 color registers and separate 6-bit (64) color palette. ECS Denise's separate nature mirrors the C128's approach.

Commodore LSI group's management rather focused on the C65's 65CE02 CPU and 256 color chipset projects.

Learning from ECS delay and conflicting with C65 mess, Amiga's Systems Engineering group created its own VLSI group who will handle AAA, AA, AA+ and Hombre.

Commodore's upper management wasn't able to combine the LSI and Systems Engineering direction.

Concerns about LSI group's competency was shown in A500's Gary chip R&D development which caused Systems Engineering group to partner with VLSI Technology Inc instead of CSG's LSI group.


4. "Super A500" was spec'ed with 1 MB VRAM as Chip RAM which is expensive while A3000 variant 2 MB Chip RAM which is crazy expensive.

IBM 8514 has 512 KB 40 ns serial access VRAM.

The balanced design is found in 3DO's MADAM i.e. Agnus role chip has access to both DRAM and VRAM.


5. Akiko's hardware C2P was born out of rebellion i.e. Commodore management is against it since they didn't care about the chunky pixel issue for the Amiga.

Commodore management banned bitplane to pixel converter. The rebels worded a requirement as speed up bitplane to whatever. Hardware C2P was inserted in by stealth.

Hardware C2P was initially created for ECS based CDTV-CR project which is recycled for CD32 project.

Commodore management applied the barebone approach for CD32 against Jeff Porter's original CD32 specs.

_________________
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:

matthey wrote: @cdimauro That 1975 8088 shown on the chart must be rare and valuable 4 years before introduction.

Probably it's a mistake: it's too early. Maybe it was the 8085, but it arrived one year after.
Quote:

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

The 68000 looks like the big boss here, for its time.
Quote:

Note: CMOS uses more transistors and area than NMOS but power draw is a fraction of NMOS.

And it was able to scale much better.

It looks like that on MOSFETs it wasn't possible to reduce resistors and capacitors (?) like transistors, hence reducing the scaling factor with new nodes. However, I have no clear information about that.
Quote:

It is difficult to find reliable information and I am unsure of some of the info. The Amiga chipset 5,000nm chip process size is about mid-70s from the data above and Wikichip.org says so too. https://en.wikichip.org/wiki/5_µm_lithography_process It wasn't good starting so far behind the tech curve but aiming higher increased costs and decreased the number of businesses that could have brought the Amiga chipset to market.

Actually you're the only one which noticed the above picture which I've added to the article, and precisely to show the huge difference between Intel's nodes timeline and how poor and far away from it was the MOS one, which crippled Commodore's chip evolution.
Quote:

If Amiga Corporation had been able to obtain financing and stay independent, the low aim likely would have looked better. Development was so much more difficult and expensive back then that a small fabless semi business was too far ahead of its time. Aiming a little more modern likely would have allowed a higher transistor budget, higher chipset clock speeds and other goodies from inception.

I agree. But that was the maximum that the Amiga Corporation was able to achieve with its financial situation, unfortunately.
Quote:

It's hard to believe the 68k Amiga was evolutionary using decade old chip technology. The original 68000 chip process was almost as outdated by the mid-1980s too. The 68000 was evolutionary itself creating the affordable personal MPU workstation market by antiquating timeshare minicomputers just half a decade earlier. That time was almost up so it was relegated to king of the 16/32 bit embedded market for another decade or more and provided a little desktop service too.

I don't get why Motorola hasn't shrunk it using better processes and allowing to reach even higher frequencies. The 68000 was one of the most used and sold processors for very long time.

@Hammer

Quote:

Hammer wrote: @cdimauro Under Henri Rubin's administration, 1. Rubin focused on add-on improvements for the Amiga OCS e.g. Amber, A2024 high res monitor, X86 bridgeboards, Unix and R&D. Henri Rubin didn't focus Amiga's core graphics upgrades.

Was him a PC guy?
Quote:

Lower ranks such as Jeff Porter was pushing for 8 bitplanes with 16 million colors and Dale Luck was pushing for 16-bit (65,536) color.

Too much at the time.
Quote:

3DO's sustained performance mode is a palettized 16-bit (65,536) color mode with 16.7 million colors palette.

My article stopped on 1990. Let's wait for the next one.
Quote:

Jeff Porter's "8 bit planes with 16 million colors" push has materialized as AA or AGA.

Chat is cheap: 16 millions, when? On 1987?
Quote:

Henri Rubin was moved to Commodore board and CDTV group before exiting from Commodore. Henri Rubin has no experience with computers. AA600 (A1200) was spec'ed by Jeff Franks from Commodore PC clone. AA600 ordered by Mehdi Ali in Feb 1992 against Bill Sydnes direction. Commodore Germany wanted the AA500 with a hard disk ASAP. A300(A600)'s PCMCIA is just Commodore's management's anti-GVP.

See above: wait for the next article. I've stopped on 1990 on this one.
Quote:

2. As a weak #metoo move, Henri Rubin ordered "monochrome hi-res Denise" and canceled Amiga Ranger R&D in mid-1986.

As I've already analyzed on my previous article, Ranger was a Mission: Impossible at the time.
Quote:

3. VGA's release around July 1987 led to "monochrome hi-res Denise" R&D wastage and moved to color high res Denise. Commodore LSI group rejected 8 color registers with shared 4096 color palette color high res Denise. Commodore LSI group's excuse is "too complex". Commodore LSI group compromised with 4 color registers and separate 6-bit (64) color palette. ECS Denise's separate nature mirrors the C128's approach. Commodore LSI group's management rather focused on the C65's 65CE02 CPU and 256 color chipset projects. Learning from ECS delay and conflicting with C65 mess, Amiga's Systems Engineering group created its own VLSI group who will handle AAA, AA, AA+ and Hombre. Commodore's upper management wasn't able to combine the LSI and Systems Engineering direction. Concerns about LSI group's competency was shown in A500's Gary chip R&D development which caused Systems Engineering group to partner with VLSI Technology Inc instead of CSG's LSI group.

It's difficult to find competent people, in general.
Quote:

4. "Super A500" was spec'ed with 1 MB VRAM as Chip RAM which is expensive while A3000 variant 2 MB Chip RAM which is crazy expensive. IBM 8514 has 512 KB 40 ns serial access VRAM.

That's a completely different market.
Quote:

The balanced design is found in 3DO's MADAM i.e. Agnus role chip has access to both DRAM and VRAM.

Let's wait for the next article.
Quote:

5. Akiko's hardware C2P was born out of rebellion i.e. Commodore management is against it since they didn't care about the chunky pixel issue for the Amiga. Commodore management banned bitplane to pixel converter.

That was a good decision.
Quote:

The rebels worded a requirement as speed up bitplane to whatever. Hardware C2P was inserted in by stealth.

They should have inserted packed/chunky mode instead of this crap.
Quote:

Hardware C2P was initially created for ECS based CDTV-CR project which is recycled for CD32 project.

Ah, nice to know. But... who created this crap?
Quote:

Commodore management applied the barebone approach for CD32 against Jeff Porter's original CD32 specs.

Another stupid decision...

@cdimauro

Quote:

@cdimauro Was him a PC guy?

Henri Rubin is a 1950s electrical engineer who doesn't have experience with computers.

Quote:

@cdimauro Chat is cheap: 16 millions, when? On 1987?

Around September 1987.

From Commodore - The Final Years
Quote:

Chipset Crisis ---------------------- Commodore’s LSI group labored through 1987 to complete the Hi-Res chipset, Denise and Agnus. The team ended up blowing past the original schedule, which called for samples in May and the first 1000 production chips in July 1987. In fact, the first samples arrived in August, and by September they were still working out bugs. George Robbins had been put in charge of overseeing the new chipset, along with the LSI engineers. Robbins felt they had done a good job taking over from the Amiga designers, (skip) By early September the West Chester engineers collectively faced an existential crisis with regards to the goals of the Hi-Res chipset. By mid-1987, they began to notice that the IBM PC and Macintosh were catching up to the Amiga. (skip) Dale Luck preferred attempting 16-bit color first, followed by a 24-bit color next generation chipset. System Plans for 1988 -------------------------------- After George Robbins completed the Amiga 500 and accompanying A520 video adapter, it was time to move on to the next project. With the prototype hi-res Agnus and Denise chips expected soon, Jeff Porter began discussing the next iteration of the A500 with Robbins. (skip) Porter specifically wanted 1000 by 800 resolution with 8 bit planes and 16 million colors—something to exceed the current competition. Of importance would be keeping the new video technology compatible with existing commercial Amiga software. All of these plans would be discussed at Commodore’s worldwide engineering meeting, scheduled for September 22, 1987 at the Embassy Suites hotel in New York. The meeting was called by Henri Rubin, and those invited included the main West Chester engineers, along with engineers from Germany and Japan.

Quote:

@cdimauro They should have inserted packed/chunky mode instead of this crap.

Context: Hardware C2P was done by the CDTV group, not by Amiga's main Systems Engineering group.

With AA chipset's refinement being frozen, Bill Sydnes (supported by Jeff Franks) kicked Jeff Porter from Amiga's Systems Engineering group and moved to the CDTV group.

I would focus on "Jeff Franks" from the Commodore PC group.

Jeff Franks replaced Jeff Porter. Jeff Franks wanted to control both Amiga and PC groups.

Jeff Franks advocated Amigas in low-end while PC for mid to high-end SKU models.

Besides Bill Sydnes, Jeff Franks is instrumental with the ECS A600 and ECS A1000Jr projects.

Quote:

@cdimauro Ah, nice to know. But... who created this crap?

CDTV group's Hedley Davis.

Do you expect Hedley Davis to single-handedly modify Denise chip with packed pixels?

You need to know Commodore's organizational structure.

From Commodore - The Final Years,
Quote:

“In the CDTV-CR there's a bitplane-to-pixel converter. I was told, ‘No, you can't do that,’ by this guy who wanted to be the boss.� Davis wanted it so badly, he decided to trick his boss. “He wanted a list of all the features that we're putting in the CDTV-CR design so that this is what we agreed to do,� says Davis. “I write a list of these nine features that we were putting in the CDTV-CR and I hand it to him and he takes it. I know he's not going to fucking read it or pay attention to it or really think about it. He's just bossing me around and just being the boss.� One of those features was the forbidden bitplane-to-pixel converter, disguised slightly. “The CDTV-CR comes out and lo and behold there is this bitplane-to-pixel converter,� recalls Davis. “He's like, ‘I can't believe you did that, it wasn't on the list. We had an agreement and you're going to be fired.’� Davis insisted the item was on the list. “He's like, ‘It wasn't on the list.’ I said get the list out. It was at like number eight and I worded it so I never used the nouns ‘bitplane to pixel converter’ but rather came up with some obtuse wording about allowing better graphics processing by minimizing processor overhead for conversion between bitplane or whatever. It was in there but I tricked him by putting it in such a way because I knew he wouldn't read it.�

CDTV-CR's bitplane-to-pixel converter, cost-reduced CD-ROM drive, and FMV MPEG1 module were reused for the AA-based CD32 project.

CDTV group wouldn't be able to modify Denise.

Last edited by Hammer on 02-Sep-2024 at 04:29 PM.
Last edited by Hammer on 02-Sep-2024 at 04:27 PM.
Last edited by Hammer on 02-Sep-2024 at 04:19 PM.
Last edited by Hammer on 02-Sep-2024 at 04:12 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:

Hammer wrote: @cdimauro Quote: @cdimauro Was him a PC guy? Henri Rubin is a 1950s electrical engineer who doesn't have experience with computers.

But at least he was a technician and his field is heavily involved on computers.
Quote:

Quote: @cdimauro Chat is cheap: 16 millions, when? On 1987? Around September 1987. From Commodore - The Final Years Quote: Chipset Crisis ---------------------- Commodore’s LSI group labored through 1987 to complete the Hi-Res chipset, Denise and Agnus. The team ended up blowing past the original schedule, which called for samples in May and the first 1000 production chips in July 1987. In fact, the first samples arrived in August, and by September they were still working out bugs. George Robbins had been put in charge of overseeing the new chipset, along with the LSI engineers. Robbins felt they had done a good job taking over from the Amiga designers, (skip) By early September the West Chester engineers collectively faced an existential crisis with regards to the goals of the Hi-Res chipset. By mid-1987, they began to notice that the IBM PC and Macintosh were catching up to the Amiga. (skip) Dale Luck preferred attempting 16-bit color first, followed by a 24-bit color next generation chipset. System Plans for 1988 -------------------------------- After George Robbins completed the Amiga 500 and accompanying A520 video adapter, it was time to move on to the next project. With the prototype hi-res Agnus and Denise chips expected soon, Jeff Porter began discussing the next iteration of the A500 with Robbins. (skip) Porter specifically wanted 1000 by 800 resolution with 8 bit planes and 16 million colors—something to exceed the current competition. Of importance would be keeping the new video technology compatible with existing commercial Amiga software. All of these plans would be discussed at Commodore’s worldwide engineering meeting, scheduled for September 22, 1987 at the Embassy Suites hotel in New York. The meeting was called by Henri Rubin, and those invited included the main West Chester engineers, along with engineers from Germany and Japan.

OK, but on September 1987 there should have been the first meeting to DISCUSS it. It's evident that it was a move to contrast Apple's Mac II and especially IBM's PS/2 line of computers which have set new standards for the mainstream graphics.

However, and after that, it required a design for the new chipset, its implementation, and tests before going to production. Read: it required some years.
Quote:

Quote: @cdimauro They should have inserted packed/chunky mode instead of this crap. Context: Hardware C2P was done by the CDTV group, not by Amiga's main Systems Engineering group. With AA chipset's refinement being frozen, Bill Sydnes (supported by Jeff Franks) kicked Jeff Porter from Amiga's Systems Engineering group and moved to the CDTV group. I would focus on "Jeff Franks" from the Commodore PC group. Jeff Franks replaced Jeff Porter. Jeff Franks wanted to control both Amiga and PC groups. Jeff Franks advocated Amigas in low-end while PC for mid to high-end SKU models. Besides Bill Sydnes, Jeff Franks is instrumental with the ECS A600 and ECS A1000Jr projects. Quote: @cdimauro Ah, nice to know. But... who created this crap? CDTV group's Hedley Davis. Do you expect Hedley Davis to single-handedly modify Denise chip with packed pixels? You need to know Commodore's organizational structure. From Commodore - The Final Years, Quote: “In the CDTV-CR there's a bitplane-to-pixel converter. I was told, ‘No, you can't do that,’ by this guy who wanted to be the boss.� Davis wanted it so badly, he decided to trick his boss. “He wanted a list of all the features that we're putting in the CDTV-CR design so that this is what we agreed to do,� says Davis. “I write a list of these nine features that we were putting in the CDTV-CR and I hand it to him and he takes it. I know he's not going to fucking read it or pay attention to it or really think about it. He's just bossing me around and just being the boss.� One of those features was the forbidden bitplane-to-pixel converter, disguised slightly. “The CDTV-CR comes out and lo and behold there is this bitplane-to-pixel converter,� recalls Davis. “He's like, ‘I can't believe you did that, it wasn't on the list. We had an agreement and you're going to be fired.’� Davis insisted the item was on the list. “He's like, ‘It wasn't on the list.’ I said get the list out. It was at like number eight and I worded it so I never used the nouns ‘bitplane to pixel converter’ but rather came up with some obtuse wording about allowing better graphics processing by minimizing processor overhead for conversion between bitplane or whatever. It was in there but I tricked him by putting it in such a way because I knew he wouldn't read it.� CDTV-CR's bitplane-to-pixel converter, cost-reduced CD-ROM drive, and FMV MPEG1 module were reused for the AA-based CD32 project. CDTV group wouldn't be able to modify Denise.

This contradicts the declarations of the CD32 technical group which has clearly stated that the idea was found during a lunch break and its design ready on the next day.

Only one thing can be true: one of the two groups is clearly lying...

@cdimauro

Quote:

cdimauro wrote: @Hammer This contradicts the declarations of the CD32 technical group which has clearly stated that the idea was found during a lunch break and its design ready on the next day. Only one thing can be true: one of the two groups is clearly lying...

A list of features is not the same as an actual chip design. The CDTV-CR does not appear to have a C2P converter in it.

Amazing that they got the machine to this stage before canning it. Just imagine what they could have produced if they had all got together to produce AGA machines from the start, instead of wasting resources on projects which weren't commercially viable (CDTV, A3000 etc.). Still, I suppose the experience gained helped them when designing the A1200 and CD32.

Even if C2P was just an idea for the CDTV-CR, at least they were thinking about it. If I was on the AGA team I would have pushed for a chunky 256 color mode, even if it was incompatible with other features and only available by banging the hardware (no OS support). VGA had it, so the Amiga desperately needed an equivalent.

I would also have looked at doing lower resolution chunky modes, such as 160 pixels which could have easily been achieved with a single bitplane superhires bitmap. 4 bytes (one 32 bit DMA fetch) would be 4 pixels. Just need a 4 way MUX to fed them into the CLUT. This would have been an ideal screen mode for Wolf 3D and Doom! For kicks I would also do a 16 color chunky mode using 4 bits per pixel and up to 320 pixel resolution. Then I would modify the Blitter slightly to make eg. color 0 (only) transparent in chunky mode - no mask required! That alone would make blitting objects 33% faster.

@cdimauro

Quote:

But at least he was a technician and his field is heavily involved on computers.

Not correct.

From Commodore - The Final Years
Quote:

Management Moves ---------------------------- 1987 The previous year, Chief Operating Officer Henri Rubin had played a crucial role in undermining Thomas Rattigan, leading to the latter’s eventual dismissal from Commodore. (skip) Henri Rubin ----------------- While working for South African company C. J. Fuchs Electrical Industries, Rubin had pioneered the Residual-current device (RCD), a safety device similar to the ground fault interrupter (GFI) that prevented workers from the dangers of electrocution. “He was the alleged inventor of the ground fault interrupter when he worked for Westinghouse in Johannesburg, South Africa,� says Commodore engineer Bill Gardei. “Miners were too often getting electrocuted by pumps used to pull water out of the mines.� RCD and GFI later came into wide use among businesses and homes. (skip) Promoted Without Being Promoted -------------------------------------------------- Following Gerard Bucas’ departure, Henri Rubin took over as VP of engineering, with Jeff Porter handling the day-to-day affairs until a replacement was found. Rubin was known for being slow to reach decisions, or not making any decisions, and generally being ineffectual. As such, he never really got around to finding a replacement, and a new status quo took shape. (skip) Productivity Mode Rubin had also instigated the A2024 “Hedley Hi-Res� monitor in early 1987. Now, with the Hi-Res chipset specs and design finalized, better business-level graphics were on the horizon with 640 by 480 non-interlaced graphics with 4 on-screen colors. Rubin instructed his Amiga programmers to start implementing these different modes into the next version of AmigaOS, which at the time was called 1.4 (it would later be renamed 2.0). Already Dale Luck had added support for the A2024 monitor. Now the programmers needed to add support for the new non-interlaced modes of the Hi-Res chipset as well. In July 1988, Rubin decided to coin a new term to describe the new video modes. “Productivity Mode was what management was asking for,� says Nesbitt. “So Productivity Mode was the code name for word processing and spreadsheets and something you could sell into an office.� Rubin became an evangelist for Productivity Mode, both within the company and outside it, while ignoring the Amiga game market. “They came up with this crazy idea to create productivity mode and refocus the Amiga on business applications, while ignoring the things that were selling well,� says Nesbitt. “Commodore just ignored the things that the computer was good at and under-invested in chasing goals that never happened.� He promised to dazzle Irving Gould at an upcoming demonstration of Productivity Mode in September.

Henri Rubin's electrical background is similar to my Dad's repairing 1970 era TVs and my Dad doesn't know how computer hardware work.

5000 units of A2024 “Hedley Hi-Res� monitor doesn't have economies of scale when competition has millions of VGA.

From Dataquest November 1989, VGA crossed more than 50 percent market share in 1989 i.e. 56%.
http://bitsavers.trailing-edge.com/components/dataquest/0005190_PC_Graphics_Chip_Sets--Product_Analysis_1989.pdf

Low-End PC Graphics Market Share by Standard Type
Estimated Worldwide History and Forecast


Total low-end PC graphic chipset shipment history and forecast
1987 = 9.2. million, VGA 16.4% market share i.e. 1.5088 million VGA.
1988 = 11.1 million, VGA 34.2% i.e. 3.79 million VGA.
1989 = 13.7 million, VGA 54.6% i.e. 7.67 million VGA.
1990 = 14.3 million, VGA 66.4% i.e. 9.50 million VGA.
1991 = 15.8 million, VGA 76.6% i.e. 12.10 million VGA.
1992 = 16.4 million, VGA 84.2% i.e. 13.81 million VGA.
1993 = 18.3 million, VGA 92.4% i.e. 16.9 million VGA.

Henri Rubin doesn't understand computer platform install base issues.

Jeff Porter argued for 8 bitplanes in 1987.

Quote:

@cdimauro This contradicts the declarations of the CD32 technical group which has clearly stated that the idea was found during a lunch break and its design ready on the next day.

There's no contradiction. CD32 evolved from CDTV-CR.

https://bigbookofamigahardware.com/bboah/product.aspx?id=1604
Quote:

If Commodore had had the budget, a new ASIC would have been built, with just the CD32 drive logic and corner turn memory coupled with the interface logic, which would have made room for a 68030 socket. The whole point was to show that Commodore engineering was healthy and had roadmap into the future, not just at the high end with Acutiator that Dave was working on, but also at the consumer end with taking the CD32 capabilities to the A1200. There were still rumors of interested buyers as late as Feb 1994. We were hoping that these kinds of projects would encourage a buyer. All Akiko did was to put 4 chips that normally take up space on a motherboard (2 Bridge chips like Gayle and Budgie, and 2 8520 CIAs), and put them in a single chip along with logic to run the CD (modified from the CD Drive logic in the CDTV-CR) and the corner-turn “chunky to planar� memory, which was a new idea to help with porting PC and Mac software to the Amiga. The “chunky to planar� logic was thought out in a lunchtime conversation between Beth Richard (system chip design), Chris Coley (board design), and Ken Dyke (software) over Subway sandwiches on a picnic table in a nearby park one day, because Ken was telling us how much of a pain it was to shuffle bits in software to port games from other platforms to the Amiga planar system. We took the idea to Hedley Davis, who was the system chip team manager and lead engineer on Akiko and he said we could go ahead with it. I showed him the “napkin sketch� of how I thought the logic would work and was planning on getting to it the next day as it was already late afternoon by that point. I came in the next morning and Hedley had completed it already, just from the sketch!

Jeff Porter's CDTV-CR date range context is during AA R&D refinement being frozen.

For Jeff Porter's CD32 project, Amiga's Systems Engineering group was involved and accepted hardware C2P as a quick fix. Hardware C2P accelerated AmigaOS 3.1's chunky pixel API. Jeff Porter tried to improve CD32's specs i.e. 8 MB RAM ($20 extra), integrated FMV (with MIPS-X SoC). Jeff Porter has to justify every component in CD32, hence upper managers have bareboned CD32.

After Mehdi Ali's Feb 1992 AA directive, AA600 was spec'ed by Jeff Frank, hence its barebone design which is missing CDTV-CR's hardware C2P.

Last edited by Hammer on 03-Sep-2024 at 04:36 AM.
Last edited by Hammer on 03-Sep-2024 at 12:59 AM.
Last edited by Hammer on 03-Sep-2024 at 12:51 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

@bhabbott

Quote:

bhabbott wrote: @cdimauro Quote: cdimauro wrote: @Hammer This contradicts the declarations of the CD32 technical group which has clearly stated that the idea was found during a lunch break and its design ready on the next day. Only one thing can be true: one of the two groups is clearly lying... A list of features is not the same as an actual chip design. The CDTV-CR does not appear to have a C2P converter in it. Amazing that they got the machine to this stage before canning it. Just imagine what they could have produced if they had all got together to produce AGA machines from the start, instead of wasting resources on projects which weren't commercially viable (CDTV, A3000 etc.). Still, I suppose the experience gained helped them when designing the A1200 and CD32. Even if C2P was just an idea for the CDTV-CR, at least they were thinking about it. If I was on the AGA team I would have pushed for a chunky 256 color mode, even if it was incompatible with other features and only available by banging the hardware (no OS support). VGA had it, so the Amiga desperately needed an equivalent. I would also have looked at doing lower resolution chunky modes, such as 160 pixels which could have easily been achieved with a single bitplane superhires bitmap. 4 bytes (one 32 bit DMA fetch) would be 4 pixels. Just need a 4 way MUX to fed them into the CLUT. This would have been an ideal screen mode for Wolf 3D and Doom! For kicks I would also do a 16 color chunky mode using 4 bits per pixel and up to 320 pixel resolution. Then I would modify the Blitter slightly to make eg. color 0 (only) transparent in chunky mode - no mask required! That alone would make blitting objects 33% faster.

http://www.bambi-amiga.co.uk/amigahistory/prototypes/cdtvcr.html

http://amiga.resource.cx/photos/photo2.pl?id=cdtv2&pg=1&res=hi&lang=en

CDTV-CR has two extra custom chips not found in other Amigas i.e. Beauty 391246-01 and Grace 391245-02 (includes Gary/Gayle role).

CDTV-CR also includes a system on a chip CSG 4510 that was developed for C65.

Last edited by Hammer on 03-Sep-2024 at 05:14 AM.
Last edited by Hammer on 03-Sep-2024 at 04:57 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

@bhabbott

Quote:

bhabbott wrote: @cdimauro Quote: cdimauro wrote: @Hammer This contradicts the declarations of the CD32 technical group which has clearly stated that the idea was found during a lunch break and its design ready on the next day. Only one thing can be true: one of the two groups is clearly lying... A list of features is not the same as an actual chip design. The CDTV-CR does not appear to have a C2P converter in it.

It wasn't a list of features: words are important, and what was reported is telling, without any doubt, that the C2P feature was actually IMPLEMENTED on the CDTV-CR.

Hardware C2P was done by the CDTV group, not by Amiga's main Systems Engineering group.
[...]
In the CDTV-CR there's a bitplane-to-pixel converter
[...]
The CDTV-CR comes out and lo and behold there is this bitplane-to-pixel converter
[...]
CDTV-CR's bitplane-to-pixel converter, cost-reduced CD-ROM drive, and FMV MPEG1 module were reused for the AA-based CD32 project.

I'm not a native speaker, as you know, but that's elementary English which anyone can understand. I've highlighted the relevant parts that prove that this logic was ALREADY IMPLEMENTED on the CDTV-CR: there's absolutely no doubt about that.
Quote:

Amazing that they got the machine to this stage before canning it. Just imagine what they could have produced if they had all got together to produce AGA machines from the start, instead of wasting resources on projects which weren't commercially viable (CDTV, A3000 etc.). Still, I suppose the experience gained helped them when designing the A1200 and CD32.

You can gain whatever experience that you like, but at the expense of delays on product lines which made the platform not competitive and finally killed the company.
Quote:

Even if C2P was just an idea for the CDTV-CR, at least they were thinking about it.

No, it was NOT an idea but a CONCRETE feature which was IMPLEMENTED: see above!
Quote:

If I was on the AGA team I would have pushed for a chunky 256 color mode, even if it was incompatible with other features and only available by banging the hardware (no OS support). VGA had it, so the Amiga desperately needed an equivalent.

Exactly, but it was useful already BEFORE the AGA, like I've reported on the article.

AND it was very easy to implement using exactly the same bitplanes/pointers logic of the chipset, as I've already explained on my older article about Akiko.
Quote:

I would also have looked at doing lower resolution chunky modes, such as 160 pixels

Please, it's 1990: we don't need another CGA/C64 mode!
Quote:

which could have easily been achieved with a single bitplane superhires bitmap.

Well, absolutely no: you don't need it. This requires too many changes in the chipset for it, whereas my solution perfectly fits on the existing chipset implementation, and works from low-res to whatever high resolution mode (having enough bandwidth, of course).
Quote:

4 bytes (one 32 bit DMA fetch) would be 4 pixels. Just need a 4 way MUX to fed them into the CLUT. This would have been an ideal screen mode for Wolf 3D and Doom! For kicks I would also do a 16 color chunky mode using 4 bits per pixel and up to 320 pixel resolution. Then I would modify the Blitter slightly to make eg. color 0 (only) transparent in chunky mode - no mask required! That alone would make blitting objects 33% faster.

25% faster. Anyway, I've covered all of this in the article: you just need to read it.

@Hammer

Quote:

Hammer wrote: @cdimauro Quote: cdimauro wrote: @Hammer This contradicts the declarations of the CD32 technical group which has clearly stated that the idea was found during a lunch break and its design ready on the next day. There's no contradiction. CD32 evolved from CDTV-CR. https://bigbookofamigahardware.com/bboah/product.aspx?id=1604 Quote: If Commodore had had the budget, a new ASIC would have been built, with just the CD32 drive logic and corner turn memory coupled with the interface logic, which would have made room for a 68030 socket. The whole point was to show that Commodore engineering was healthy and had roadmap into the future, not just at the high end with Acutiator that Dave was working on, but also at the consumer end with taking the CD32 capabilities to the A1200. There were still rumors of interested buyers as late as Feb 1994. We were hoping that these kinds of projects would encourage a buyer. All Akiko did was to put 4 chips that normally take up space on a motherboard (2 Bridge chips like Gayle and Budgie, and 2 8520 CIAs), and put them in a single chip along with logic to run the CD (modified from the CD Drive logic in the CDTV-CR) and the corner-turn “chunky to planar� memory, which was a new idea to help with porting PC and Mac software to the Amiga. The “chunky to planar� logic was thought out in a lunchtime conversation between Beth Richard (system chip design), Chris Coley (board design), and Ken Dyke (software) over Subway sandwiches on a picnic table in a nearby park one day, because Ken was telling us how much of a pain it was to shuffle bits in software to port games from other platforms to the Amiga planar system. We took the idea to Hedley Davis, who was the system chip team manager and lead engineer on Akiko and he said we could go ahead with it. I showed him the “napkin sketch� of how I thought the logic would work and was planning on getting to it the next day as it was already late afternoon by that point. I came in the next morning and Hedley had completed it already, just from the sketch! Jeff Porter's CDTV-CR date range context is during AA R&D refinement being frozen. For Jeff Porter's CD32 project, Amiga's Systems Engineering group was involved and accepted hardware C2P as a quick fix. Hardware C2P accelerated AmigaOS 3.1's chunky pixel API. Jeff Porter tried to improve CD32's specs i.e. 8 MB RAM ($20 extra), integrated FMV (with MIPS-X SoC). Jeff Porter has to justify every component in CD32, hence upper managers have bareboned CD32. After Mehdi Ali's Feb 1992 AA directive, AA600 was spec'ed by Jeff Frank, hence its barebone design which is missing CDTV-CR's hardware C2P.

See above my reply to Bruce: the C2P feature was already implemented on the CDTV-CR, and it was NOT just an idea coming from Davis. At least from what YOU have previously reported.

What you've reported now comes from the CD32 development, and your text clearly says that it is a NEW IDEA.

Hence: the clear contradiction.

Words are important, and there's no doubt that the CD32's C2P was NOT the same of the CDTV-CR, but something NEW.

@cdimauro

From Commodore - The FInal Years
Quote:

AA First Prototypes Back in September 1989, George Robbins proposed an intermediate level Amiga chipset, called Pandora, to bridge the gap until AAA appeared. The chipset, consisting of a graphics chip called Lisa and an improved Agnus called Alice, was supposed to be ready in early 1990. Rubin had given the go ahead and six chip engineers had been working full time under the project leader, Bob Raible. (skip) As work progressed on the graphics chip, Lisa, it became clear that the timeline was too ambitious. Jeff Porter set a more realistic expectation of late 1990. As Porter predicted, CSG produced the first prototype chips by late November 1990 and testing began. By early December, the team felt the Lisa chip would be delivered ahead of schedule. All but the color table was working, a problem Raible felt he could overcome with a hack. (skip) The AA chips continued to be revised and tested through early 1991 until they were good enough to use in the A1000 Plus and A3000 Plus prototypes. Dave Haynie managed to boot up his A3000 Plus with AmigaOS and the AA chipset in February 1991. (skip) By March 27, when the tested AA chips were ready, the list of stable features was impressive. AA could display 256 colors from a palette of 16 million colors. It could theoretically play 24-bit digital video (although presently it could only display 8-bit video) due to a four times increase in bandwidth (and using the digital-to-analog converter chip in the A3000 Plus). It could use 64-bit sprites, which could now be controlled in the border areas. It had the aforementioned new 8-bit HAM mode, called Super HAM. And finally, the scan doubling and deinterlacing hardware, formerly on Amber, was now handled right on the chip. There were also nine features the engineers discussed in October 1989 which had not yet been implemented, including: 16-bit processor support, 8 MB chip RAM, 16- and 24-bit color modes, faster HAM modes, and a simple integrated digital-to-analog converter (for the low-end Amigas). They also had a list of 10 other features they wanted to add to the AA chipset, primarily in Paula, including: high density floppy support, 16 bit sound, audio input support, (skip)

AA's R&D refinement was frozen for more than six months.

_________________
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

@cdimauro

Quote:

Words are important, and there's no doubt that the CD32's C2P was NOT the same of the CDTV-CR, but something NEW.

AA has 8 bitplanes while ECS has 5 or 6 bitplanes, hence it's not the same.

Hedley Davis is responsible for both hardware C2P in CDTV-CR and CD32's Akiko.

Jeff Franks is responsible for Amiga Systems Engineering Group's A600 ECS and AA600/A1200.

Jeff Porter is responsible for the multimedia group's CDTV-CR ECS and CD32 AA.

Last edited by Hammer on 03-Sep-2024 at 05:49 AM.
Last edited by Hammer on 03-Sep-2024 at 05:47 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

@Hammer

Quote:

Hammer wrote: @cdimauro From Commodore - The FInal Years Quote: AA First Prototypes Back in September 1989, George Robbins proposed an intermediate level Amiga chipset, called Pandora, to bridge the gap until AAA appeared. The chipset, consisting of a graphics chip called Lisa and an improved Agnus called Alice, was supposed to be ready in early 1990. Rubin had given the go ahead and six chip engineers had been working full time under the project leader, Bob Raible. (skip) As work progressed on the graphics chip, Lisa, it became clear that the timeline was too ambitious. Jeff Porter set a more realistic expectation of late 1990. As Porter predicted, CSG produced the first prototype chips by late November 1990 and testing began. By early December, the team felt the Lisa chip would be delivered ahead of schedule. All but the color table was working, a problem Raible felt he could overcome with a hack. (skip) The AA chips continued to be revised and tested through early 1991 until they were good enough to use in the A1000 Plus and A3000 Plus prototypes. Dave Haynie managed to boot up his A3000 Plus with AmigaOS and the AA chipset in February 1991. (skip) By March 27, when the tested AA chips were ready, the list of stable features was impressive. AA could display 256 colors from a palette of 16 million colors. It could theoretically play 24-bit digital video (although presently it could only display 8-bit video) due to a four times increase in bandwidth (and using the digital-to-analog converter chip in the A3000 Plus). It could use 64-bit sprites, which could now be controlled in the border areas. It had the aforementioned new 8-bit HAM mode, called Super HAM. And finally, the scan doubling and deinterlacing hardware, formerly on Amber, was now handled right on the chip. There were also nine features the engineers discussed in October 1989 which had not yet been implemented, including: 16-bit processor support, 8 MB chip RAM, 16- and 24-bit color modes, faster HAM modes, and a simple integrated digital-to-analog converter (for the low-end Amigas). They also had a list of 10 other features they wanted to add to the AA chipset, primarily in Paula, including: high density floppy support, 16 bit sound, audio input support, (skip) AA's R&D refinement was frozen for more than six months.

And? What has this to do with my new article and the actual discussions?

@Hammer

Quote:

Hammer wrote: @cdimauro Quote: Words are important, and there's no doubt that the CD32's C2P was NOT the same of the CDTV-CR, but something NEW. AA has 8 bitplanes while ECS has 5 or 6 bitplanes, hence it's not the same.

That's a very minor change on the C2P logic / implementation.

Certainly it does NOT require a NEW idea...
Quote:

Hedley Davis is responsible for both hardware C2P in CDTV-CR and CD32's Akiko. Jeff Franks is responsible for A600 ECS and AA600/A1200. Jeff Porter is responsible for CDTV-CR ECS and CD32 AA.

Whatever: see above.

@cdimauro

Quote:

cdimauro wrote: And? What has this to do with my new article and the actual discussions?

AA is an evolution from the existing 16-bit ECS.

AA is a 32-bit/16-bit hybrid. AA's "8 bitplanes" argument from Jeff Porter has existed since 1987.

Jeff Porter can argue, but he doesn't have Henri Rubin's "hot seat".

_________________
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

@cdimauro

Quote:

cdimauro wrote: That's a very minor change on the C2P logic / implementation. Certainly it does NOT require a NEW idea...

That's your problem. CDTV-CR's hardware C2P wasn't supposed to exist.

Last edited by Hammer on 03-Sep-2024 at 05:59 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

@Hammer

Quote:

Hammer wrote: @cdimauro Quote: cdimauro wrote: And? What has this to do with my new article and the actual discussions? AA is an evolution from the existing 16-bit ECS. AA is a 32-bit/16-bit hybrid.

It's just 16 bit: there are no 32 bits.
Quote:

AA's "8 bitplanes" argument from Jeff Porter has existed since 1987.

Which was just a discussion.
Quote:

Jeff Porter can argue, but he doesn't have Henri Rubin's "hot seat".

Neither the time machine: those were just ideas yet to be discussed. THEN they need to be finalized.

@cdimauro

Quote:

Neither the time machine: those were just ideas yet to be discussed. THEN they need to be finalized.

Again, from Commodore - The Final Years,
Quote:

AA First Prototypes Back in September 1989, George Robbins proposed an intermediate level Amiga chipset, called Pandora, to bridge the gap until AAA appeared. The chipset, consisting of a graphics chip called Lisa and an improved Agnus called Alice, was supposed to be ready in early 1990. Rubin had given the go ahead and six chip engineers had been working full time under the project leader, Bob Raible. (skip) As work progressed on the graphics chip, Lisa, it became clear that the timeline was too ambitious. Jeff Porter set a more realistic expectation of late 1990. As Porter predicted, CSG produced the first prototype chips by late November 1990 and testing began. By early December, the team felt the Lisa chip would be delivered ahead of schedule. All but the color table was working, a problem Raible felt he could overcome with a hack. (skip) The AA chips continued to be revised and tested through early 1991 until they were good enough to use in the A1000 Plus and A3000 Plus prototypes. Dave Haynie managed to boot up his A3000 Plus with AmigaOS and the AA chipset in February 1991. (skip) By March 27, when the tested AA chips were ready, the list of stable features was impressive. AA could display 256 colors from a palette of 16 million colors. It could theoretically play 24-bit digital video (although presently it could only display 8-bit video) due to a four times increase in bandwidth (and using the digital-to-analog converter chip in the A3000 Plus). It could use 64-bit sprites, which could now be controlled in the border areas. It had the aforementioned new 8-bit HAM mode, called Super HAM. And finally, the scan doubling and deinterlacing hardware, formerly on Amber, was now handled right on the chip. There were also nine features the engineers discussed in October 1989 which had not yet been implemented, including: 16-bit processor support, 8 MB chip RAM, 16- and 24-bit color modes, faster HAM modes, and a simple integrated digital-to-analog converter (for the low-end Amigas). They also had a list of 10 other features they wanted to add to the AA chipset, primarily in Paula, including: high density floppy support, 16 bit sound, audio input support, (skip)

The key words are in bold letters i.e. "Rubin had given the go ahead".

Last edited by Hammer on 03-Sep-2024 at 06:02 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