@ppcamiga1

Quote:

ppcamiga1 wrote: di mauro stop trolling start working on aros on zune

PARROT MODE ON. Then:

I reveal you another secret: YOU are the only one here which put it as The goal to have.

I reveal you also another secret: AROS is open source, so anyone can get the source and contribute.

Now, I reveal you elementary logic: since YOU are the only one interested on the above, then YOU can pick AROS' sources and add what YOU like.


Besides that, I've to decline, because I am already busy:



Would you like to join me and help me bury the PowerPCs?

@kolla

Quote:

kolla wrote: @cdimauro Quote: cdimauro wrote: @kolla That's the reason why it survived at least 'til 94: with your proposals it would have broken well before... What have I proposed? I’ve merely pointed out that looking at adding DSPs did make sense at the time,

That's very high questionable.
Quote:

and that it could have been useful for those of us who aren’t solely games focused.

Which means: a very small niche.
Quote:

This whole alternative reality thing of what they should and shouldn’t have done is all on you.

Well, you did exactly the same here. With the difference that you're advocating for an alternative reality which was much worse, since you want to have Commodore computers for small niches, which haven't sold so much in its history.

How can a company like Commodore would have survived with even more expensive computers to be produced with a so small audience?

cdimauro Quote:

The problem is that we're really too late to think about reviving the platform. Even the A500 Mini hasn't sold so much (I think it was 150k units?) considered that 5.5 million Amigas were sold.

THEA500 Mini is a not so cheap nostalgic Amiga toy with less "hardware" value than a $15 USD RPi Zero 2 W.

OneTimer1 Quote:

The 68060 wan't available in the early 90ies, according to Wikipedia is wasn't in production before 1994. According to you it cost $308 that would have increased the price of an A1200 about $600, It would have needed FastRAM adding easily $100 again.

According to the Microprocessor Report article I linked, the 68060@50MHz was introduced in 1994 at $308 USD in "1000s" quantities. CBM usually had a significant discount and the price of the 68060 would have dropped from there. CBM licensing and producing of the 68060 could have lowered prices further and potentially improved value if they could have clocked up the 68060. A 68060&AA+ SoC could have likely been produced for minimal additional cost compared to producing a 68060 chip. The integration advantages I posted were from CBM documentation.

OneTimer1 Quote:

Your ideas would have nearly doubled the price of an A1200, that is way off topic. A cheap improved Amiga would have been needed to prevent C= bankruptcy, that happened 1995. AGA would have been needed around 1990 and a compatible improvement around 1992 and not with a CPU that wasn't available.

CBM's ideas, plans, time line from linked source documentation in my last two posts. My organization and analysis (CBM docs are scattered and disorganized). Definitely on topic as the low end 16/32 bit Amiga AA+ chipset was planned for the SoC and not the high end 32/64 bit AAA chipset. AA+ was originally planned by CBM to be in available systems in 1994. The low power fully CMOS and modular AA+ chipset would have likely included a 68EC030@28MHz (CD32+ for example) but AA+ was designed to operate at 57MHz using a 1000nm fab process. The AA+ chipset was still designed to operate with a 16-bit 68000 and could have been used for low cost toys (not unlike Jeri Ellsworth Amiga chip later) and very low cost A200 and A300 systems. The low end Amiga market was partially saturated and Amiga users wanted upgraded hardware. However, a 68000-68040 CPU chip could not operate at 57MHz, without overclocking, to fully take advantage of AA+. The 68000-68030 lack pipelining to clock them up without moving to more expensive processes while the 68040 did not have enough low power features and ran too hot to clock up over 40MHz. The modular, low power and 32-bit data bus 68060, like AA+, was also due in 1994 with a deep 8-stage pipeline that was very conservatively clocked. The CBM docs do not say what 68k CPU was to be used in the SoC which was due in 1995 to 2000 depending on the documentation. AA+ was designed to be modular and flexible so multiple 68k CPU cores could have been licensed and used. I was analyzing the mid to high end of the low end Amiga chipset that would have been the better value upgrade from AGA we didn't get from CBM. While 68060&AA+@57MHz SoC Amigas likely would have been mid-priced in the mid-1990s, they likely would have been low priced by the late 1990s with 68060+&AA+&3D@114MHz at the mid to high end, with die shrinks as necessary. More than likely, AAA would have been abandoned as the high end chipset with worthwhile features ported back to more practical AA+ successors. Officially, AAA was still the goal even for mid-priced hardware but there were questions of feasibility with unnecessary 2D features and display modes.

https://ia801903.us.archive.org/22/items/amiga-chip-set-strategy/Amiga_Chip_Set_Strategy.pdf Quote:

Note: Combining a LC Lisa and Alice & Paula combo results in a device count of 146,017 transistors and would easily fit in a 208 FQFP package. From a total device count perspective this chip has approximately the same devices as the AAA Monica chip! The ultimate low end system would be this Super Amiga chip and the super I/O device with integrated 68k core.

I don't like the LC Lisa idea which I believe is unnecessary but the point that AA+ was much smaller and more practical than AAA was already obvious. Transistors were already getting cheap in the 1990s but the smaller AA+ transistor budget left room for a better CPU with larger caches and a 3D core.

CBM upper management could only see the current market and lacked vision. They allowed only the minimum Amiga chipset enhancements and patch work to address current problems rather than future problems resulting in the Amiga chipset lagging the market. Good examples include the monochrome Denise upgrade, flicker fixer, AGA non-upgraded blitter and Paula, Akiko chunky, etc. They decapitated the leadership of Jay Miner who was a visionary but not perfect.

https://archive.org/download/commodorestrategicplan19851987/Commodore_Strategic_Plan_1985-1987.pdf Quote:

Graphics Competition Steve Mayer, the president of Take One and the designer of the T-chip, was ex-head of Warner Labs and co-designer, with Jay Miner at Atari, of the Atari VCS and 800 chipsets, and enjoys a similar industry reputation. The reason the T-chip is important is that it uses a pixel address approach rather than the bit-plane approach of the Amiga chipset. Jay Miner himself has said, very confidently, that the bit-plane approach is essentially a dead-end, and that if he had to do the Amiga chipset over, he would do a pixel address design. This means that although there are some obvious enhancements to the Amiga chipset, it cannot easily do what the next generation of graphics/video chips are required to do, especially for consumer and vertical market applications. A hybrid pixel address/bit plane chipset is apparently possible but difficult, and raised many complex issues of compatibility for existing Amiga software. (A report of these issues by Miner has unfortunately been delayed several weeks.) ... The 640 x 400 non-interlaced of the Amiga chipset will not be available until late fall '86, although a partial upgrade may be possible in time for Ranger. Added features and performance enhancements will appear in mid-87. Miner feels that the pixel address chipset, which is the direction all display technology is moving, and which is necessarily completely incompatible with the Amiga, will take three years. (For all these chip designs, he will need additional staffing.)

AA+ was finally the Amiga chipset mostly back on track with chunky including bit-plane backward compatibility, upgraded chip mem bandwidth, programmable high resolution non-interlaced display, upgraded blitter, upgraded audio, etc. The problems were known in 1985 (from docs above), due to be fixed in 1987-1988 but never made it into production by 1994, likely due to CBM management delays and then not having finances to move ahead with AA+ chipset testing and production. The AA+ specs and plans indicate that AA+ chipset cores may exist though. I guess at this point it would just be another enhanced AGA chipset (at least 4 enhanced AGA chipset cores?) regardless of schematics or HDL code. There are likely powers in the Amiga community that would like to bury the AA+ chipset core (and 68k cores) as much as CBM buried the Ranger chipset. A real 68k SoC is still the way to affordable Amiga hardware if there is adequate demand for mass production. A 68060+&AA+&3D SoC on more modern silicon would be comparable to a RPi 3 using a SoC today (the original RPi had a less advanced CPU than the the 68060).

P.S. It's frustrating when people comment and criticize without reading sourced CBM docs.

Last edited by matthey on 26-Sep-2024 at 07:41 PM.
Last edited by matthey on 26-Sep-2024 at 06:13 PM.
Last edited by matthey on 26-Sep-2024 at 06:02 PM.

@cdimauro

Quote:

anyone can get the source and contribute

Nope, not anyone who’ve had a glimpse of the “original� Amiga OS source code…

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@cdimauro

Quote:

Well, you did exactly the same here. With the difference that you're advocating for an alternative reality which was much worse, since you want to have Commodore computers for small niches, which haven't sold so much in its history.

Nonsense, I have _not_ done exactly the same, I have just defended what actually did happen. I am not suggesting that CBM should have done this and that, I’m just saying that it did make sense to look at DSP - as they actually did - as it would have been useful for certain things. I am not judging what was done, but you are.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@matthey

Quote:

matthey wrote: cdimauro Quote: The problem is that we're really too late to think about reviving the platform. Even the A500 Mini hasn't sold so much (I think it was 150k units?) considered that 5.5 million Amigas were sold. THEA500 Mini is a not so cheap nostalgic Amiga toy with less "hardware" value than a $15 USD RPi Zero 2 W.

In absolute terms, yes, but the RPi does not take the Amiga ecosystem with it (software: OS, some applications, some gams). That's the added value here.
Quote:

Quote: https://ia801903.us.archive.org/22/items/amiga-chip-set-strategy/Amiga_Chip_Set_Strategy.pdf [quote] Note: Combining a LC Lisa and Alice & Paula combo results in a device count of 146,017 transistors and would easily fit in a 208 FQFP package. From a total device count perspective this chip has approximately the same devices as the AAA Monica chip! The ultimate low end system would be this Super Amiga chip and the super I/O device with integrated 68k core. I don't like the LC Lisa idea which I believe is unnecessary but the point that AA+ was much smaller and more practical than AAA was already obvious. Transistors were already getting cheap in the 1990s

Especially if you consider that Commodore was moving from its MOS to external partners, which were using the much better CMOS process.
Quote:

but the smaller AA+ transistor budget left room for a better CPU with larger caches and a 3D core. CBM upper management could only see the current market and lacked vision.

I'd like to add the technical team as well on that.
Quote:

They allowed only the minimum Amiga chipset enhancements and patch work to address current problems rather than future problems resulting in the Amiga chipset lagging the market. Good examples include the monochrome Denise upgrade, flicker fixer, AGA non-upgraded blitter and Paula, Akiko chunky, etc. They decapitated the leadership of Jay Miner who was a visionary but not perfect.

Exactly, but all of this comes also from the technical side of the company, with engineers which were like hackers and putting ugly patches everywhere to get the job done in a quick and dirty way.
Quote:

P.S. It's frustrating when people comment and criticize without reading sourced CBM docs.

Right. But such documents show also the lack of vision of the whole Commodore people.

@kolla

Quote:

kolla wrote: @cdimauro Quote: anyone can get the source and contribute Nope, not anyone who’ve had a glimpse of the “original� Amiga OS source code…

Once the original source code is public, this doesn't matter. In the worst case, those people can steer the development providing suggestions.

But if Hyperion loses the current legal battle, then the situation changes and NDAs would be void.
Quote:

kolla wrote: @cdimauro Quote: Well, you did exactly the same here. With the difference that you're advocating for an alternative reality which was much worse, since you want to have Commodore computers for small niches, which haven't sold so much in its history. Nonsense, I have _not_ done exactly the same, I have just defended what actually did happen. I am not suggesting that CBM should have done this and that, I’m just saying that it did make sense to look at DSP - as they actually did - as it would have been useful for certain things.

That's too abstract: you can put whatever you want and you'll find a way to use it, but it doesn't mean that it would make sense.

As usual, you should see the market of reference and draw your conclusions if this makes sense or not.

And the DSP definitely didn't make sense, according to it. Especially considering that the AAA was already embedding 8 x 16-bit, CD-quality, audio channels, which was the primary reason for which the DSP was looked at (according to the engineer's technical documents on that, and the interviews).
Quote:

I am not judging what was done, but you are.

Yes, because I consider the whole context, and I don't stop only to: "it could be cool to add this".