Having completed our view of the Amiga's 16-bit platform, let us move on to the 32-bit platform (which, however, lays the foundation for possible future evolutions).

English: https://www.appuntidigitali.it/23498/missed-opportunities-to-improve-the-amiga-chipset-7-the-alternative-of-32-bit-innovations/

Dopo aver completato la visione della piattaforma a 16 bit dell'Amiga, passiamo a quella a 32 bit (che, comunque, pone la basi per le possibili, future, evoluzioni).

Italian: https://www.appuntidigitali.it/23356/le-occasioni-mancate-per-migliorare-il-chipset-dellamiga-7-lalternativa-delle-innovazioni-a-32-bit/

@cdimauro

Nice read, good summary of "what IFs" ...

I wonder how do you find Apollo Core`s SAGA chipset concept of overcoming most of mentioned
possible roads to go?
http://www.apollo-core.com/sagadoc/
https://www.scribd.com/document/539527119/Sprites

_________________
Future Acube and MOS supporter, fi di good, nothing fi di unprofessionals. Learn it harder way!

@vox

Quote:

vox wrote: @cdimauro Nice read, good summary of "what IFs" ... I wonder how do you find Apollo Core`s SAGA chipset concept of overcoming most of mentioned possible roads to go? http://www.apollo-core.com/sagadoc/ https://www.scribd.com/document/539527119/Sprites

Thanks. Unfortunately I can't access the second link and it seems that there are no other copies of the PDF around, so it's very very difficult to see how it worked and make the necessary evaluations.

I can tell you, from the first link, that the audio partially matches what I've proposed, but there are deep differences.
At first sight, it looks like the philosophy was: "duplicate what's need from AGA". Which is far away from my vision and my proposals.

The only thing that partially matches is expanding the volume from 6 + 1 bits to 8 bits, but I assume that it's just doing it the way that channels 0-3 and 1-2 worked on the Amiga, which is a bit weird, whereas I've proposed a more simple and general-purpose way of handling the outputs in stereo format (and going way further to support quadraphonic and 7.1).

In short: it looks like a patchwork coming from the already very bad AGA implementation. However, and as I've said, I'm not able to make other guesses because I lack the documentation, but from the registers list my forecast is that it would confirm my opinion.

cdimauro Quote:

In short: it looks like a patchwork coming from the already very bad AGA implementation. However, and as I've said, I'm not able to make other guesses because I lack the documentation, but from the registers list my forecast is that it would confirm my opinion.

Thomas may be a perfectionist, Jens is a very neat programmer and then there is Gunnar. Sigh.

Interesting article. HP fabbed Amiga chip should be Lisa not Alice near the bottom of your article?

https://www.appuntidigitali.it/23498/missed-opportunities-to-improve-the-amiga-chipset-7-the-alternative-of-32-bit-innovations/ Quote:

Some may wonder why the whole system continues to run at 14Mhz and not 28Mhz, for example, since it would have been possible to raise the frequencies as well, two years after the previous one. The answer is that Commodore had to take into account its target market, which is the low cost/mainstream market, so it could not overdo it. In addition, continuing to rely on its MOS foundries was a big problem in terms of scalability (also in frequency), which in fact forced it to turn to HP for the implementation of the AGA’s most complicated chip (Alice: the video controller).

@cdimauro

Thanks. I thoroughly enjoyed reading this well constructed piece.

Afterwards, my mind went to that scene in Apollo 13 when all the engineers are vocally problem solving, and that one engineer chimes in over the din and delivers his "Power is Everything" clarion call.
https://youtube.com/clip/UgkxMHFwhj90xkgudPdbiP_fZExObTeJZPe6?si=92FZZpuvMiVmRhbh

I guess in the Commodore engineering room they didn't have that person to come in and proclaim "Bus bandwidth is everything", or maybe there was an they were ignored at their eventual peril.

_________________
All the way, with 68k

agami Quote:

I guess in the Commodore engineering room they didn't have that person to come in and proclaim "Bus bandwidth is everything", or maybe there was an they were ignored at their eventual peril.

Jay Miner is the CBM engineer who pretty much pointed out the importance of memory bandwidth when he designed the Ranger chipset using VRAM in 1986 without being asked to. CBM showed how important they thought memory bandwidth was by ignoring him and waiting until 1992 to finally release AGA increasing the bandwidth but only for part of the chipset. The perspective of the Amiga would have been completely different if the Amiga had a high end graphics workstation market. All CBM cared about was a low cost low margin C64 replacement.

@matthey

Quote:

matthey wrote: cdimauro Quote: In short: it looks like a patchwork coming from the already very bad AGA implementation. However, and as I've said, I'm not able to make other guesses because I lack the documentation, but from the registers list my forecast is that it would confirm my opinion. Thomas may be a perfectionist, Jens is a very neat programmer and then there is Gunnar. Sigh.

Quote:

Interesting article. HP fabbed Amiga chip should be Lisa not Alice near the bottom of your article? https://www.appuntidigitali.it/23498/missed-opportunities-to-improve-the-amiga-chipset-7-the-alternative-of-32-bit-innovations/ Quote: Some may wonder why the whole system continues to run at 14Mhz and not 28Mhz, for example, since it would have been possible to raise the frequencies as well, two years after the previous one. The answer is that Commodore had to take into account its target market, which is the low cost/mainstream market, so it could not overdo it. In addition, continuing to rely on its MOS foundries was a big problem in terms of scalability (also in frequency), which in fact forced it to turn to HP for the implementation of the AGA’s most complicated chip (Alice: the video controller).

Oh, right. It was a lapsus. Just fixed. Thanks!

@agami

Quote:

agami wrote: @cdimauro Thanks. I thoroughly enjoyed reading this well constructed piece.

Quote:

Afterwards, my mind went to that scene in Apollo 13 when all the engineers are vocally problem solving, and that one engineer chimes in over the din and delivers his "Power is Everything" clarion call. https://youtube.com/clip/UgkxMHFwhj90xkgudPdbiP_fZExObTeJZPe6?si=92FZZpuvMiVmRhbh

Great scene. I love this film also because of this: it shows how important is having the right vision. Which unfortunately Commodore engineers which replaced the original team lacked.
Quote:

I guess in the Commodore engineering room they didn't have that person to come in and proclaim "Bus bandwidth is everything", or maybe there was an they were ignored at their eventual peril.

From what Hammer reported, it was The Land of Confusion there: they have very different ideas on how to evolve the platforms. And they spent so much time on discussions.

Amiga, as it was designed, wasn't all about the bandwidth. It certainly is the most important factor to consider (that's what they had), but even more important was... how to use it to solve specific technical problems. And here you need vision...

@matthey

Quote:

matthey wrote: agami Quote: I guess in the Commodore engineering room they didn't have that person to come in and proclaim "Bus bandwidth is everything", or maybe there was an they were ignored at their eventual peril. Jay Miner is the CBM engineer who pretty much pointed out the importance of memory bandwidth when he designed the Ranger chipset using VRAM in 1986 without being asked to.

Exactly, but his problem was that... he went to ahead with his design, requesting too much at the time.

Ranger would have been a nice project for addressing workstations, but that wasn't Commodore's market.

It could have been ALSO, if the management would have had the right vision (and put the money there, of course) to develop a chipset which was able to address both the game and workstation market, in a scalable way. Which, BTW, it was what they were looking for with the Commodore 900!!!

Their ineptitude is unbelievable...
Quote:

CBM showed how important they thought memory bandwidth was by ignoring him and waiting until 1992 to finally release AGA increasing the bandwidth but only for part of the chipset. The perspective of the Amiga would have been completely different if the Amiga had a high end graphics workstation market. All CBM cared about was a low cost low margin C64 replacement.

*

cdimauro Quote:

Exactly, but his problem was that... he went to ahead with his design, requesting too much at the time. Ranger would have been a nice project for addressing workstations, but that wasn't Commodore's market.

CBM chased the old high resolution workstation market with ECS only to find the market was moving on to high resolution with color which Ranger would have given the Amiga years earlier. Then this high end with color workstation market moved to the high end desktop while CBM was playing follow the leader to catch up. The Amiga started in the lead but rather than continue to lead, as Jay planned, CBM decided to be reactive tech followers. The Amiga needed a high end chipset which could later become the low end chipset when prices came down.

cdimauro Quote:

It could have been ALSO, if the management would have had the right vision (and put the money there, of course) to develop a chipset which was able to address both the game and workstation market, in a scalable way. Which, BTW, it was what they were looking for with the Commodore 900!!!

Even with scalable features like supporting VRAM and DRAM, a high end and low end chipset may have been desirable. Supposedly high end big box Amigas with the low end chipset lacked value. Graphics cards were too expensive to improve the value as could be seen with the A3000UX with A2410 graphics card. The value advantage of using the Amiga for Unix was gone.