Sign of life of Apollo FPGA project

[matthey]

Gunnar has not promised anything, stop using that word! Besides, he is not the one doing the hardware, his hope is that Apollo will become defacto standard CPU core for Amiga, and that others will license his core for their acc card products.

From my perspective, Gunnar has promised nothing. He has stated that he would like to and plans to bring out new accelerators for different Amiga models. Gunnar has been known to change his mind and plans though.

even though im rather for open source (for security reasons) i would have nothing against if gunnar succeeds with his core even if others would have to license it. i have my doubts about extensions and establishing new standards, but thats another matter.

A single standard is good but one man making all the decisions for it is not. I would hope that those licensing the product would receive the source code with some ability to make changes.

[matthey]

people behind apollo core, be it gunnar or whomever, are making their choices, and thats their own right. if these choices will become standard depends on many factors, but also on social dynamics similar to what you might thought of as "free market" or "delivery contra demand". i thought this is sympathetic to you, since if the standard occures to be good it may be successful, and then the authors dont have much interest to keep it back. to me its strange that exactly you are calling here for "state control"

Gunnar is free to make his choices even though they may sabotage his own project and plans. IMO, Gunnar is overbearing and overwhelms other people's opinions in his "teams" which interferes with their choices. It is his right when he created the team but it is also arrogant and selfish.

im fine with this attitude. i consider this flexible. remember the long quarrel about movep and that it finally got implemented without any further discussion. questioning and reconsidering is genuinely positive, and people who appear stubborn, may not be really that much as they appear.

Gunnar finally took meynaf's and my advice in several areas (especially about compatibility) after he discredited our arguments and ignored our questions. When Gunnar didn't respect us enough, I suggested expanding the team with some other people (well respected compiler maintainers and AmigaOS developers) so a consensus could be formed but Gunnar's opinion seems to be that the processor designers by themselves (him) have privileged authority over all decisions. Gunnar's teams involve and expect loyal subordination instead of working together as diverse respected peers. Maybe the project will succeed due to personal efforts but it is unlikely to be as successful as a real "team" effort.

I am not worried, the grand plan will never happen, at "best" it just cause yet another split in the community, and in time more compatible open source cores will take over anyways.

Other open source cores will likely become more advanced and faster but not to the level of the Apollo core. The fastest FPGA Arcade and Mist cores (based off the TG68) will likely never exceed 68040 performance while the Apollo core has already exceeded 68060 performance and approaches the performance of the best ColdFire hard processors in an affordable FPGA of Majsta's latest accelerator. I have every reason to believe that the Apollo core technology is real and is a unique opportunity for the Amiga to become relevant and affordable again. Leadership and investment are lacking in the "grand plan" so you are probably correct in your pessimism.

And I repeat - all we have is _talk_ that this and that is implemented - or rather, that it is _decided_ to _become_ implemented. So far I have yet to see any concrete evidence of anything when it comes to actual compatibility and the features claimed.

Frustration leads to pessimism, pessimism leads to anger, anger leads to the dark side. You are obviously an intelligent person so why are you so unjustifiably negative?

[matthey]

You might want to word that more carefully.

The 68K implementation running on the Mist/FPGAA of _today_ might stop at 040.
That same core can be re-targeted for, say, a Stratix 10 chip (we _are_ used to pay through the nose for our hw anyway) and it would probably be 2-5 times faster, mostly as-is. No, there are no FPGA designs out there with that chip on, but why should the Mist/FPGAA be the end of the evolutionary line?

I was assuming that the next generation FPGA hardware would get new names but I suppose significant upgrades with the same name are possible. You are correct that there are more expensive FPGAs available which could make the Apollo core faster than the fastest ColdFire hard CPU. The Stratix is pretty expensive but there are some medium priced FPGAs which would double the performance and still come out at less than half the price of new AmigaOS 4 hardware.

As soon as there is an 060@300-400MHz FPGA with full AGA it's time to switch :-)

Two more weeks .

[matthey]

Why? Really, why do you say so? Are you saying Gunnar is some sort of uniq wizard with supernatural abilities?

Gunnar, Jens and Chris have a very good understanding of both processor design and FPGA programming. Granted, most processor designers today pretty much have to have good FPGA programming skills as this is how processors are made but the good ones make huge salaries. These guys worked in the field (for IBM) before they decided to create their own processor design. Employees who split from larger companies to innovate outside of the usually more restrictive big business atmosphere is where most successful small businesses come from. It just so happens that these guys are Amiga and 68k fans who want to do something with the Amiga and 68k. Thomas Hirsch, the Natami designer, is also an Amiga fan who is a professional hardware designer. These guys want to work on the Amiga and 68k and I believe would be willing to work for less to do what they love and help bring back the Amiga from the edge of extinction. I know as I worked for free helping to design and document a new 68k ISA, made a 68k code analyzer from a disassembler and made suggestions based on my assembler and compiler experience (Gunnar is using several of my ideas and suggestions). Maybe it is too late for the Amiga. All the big guys want to continue a dead end route making what is left of the Amiga from PC hardware for the classes instead of the masses.

So what? FPGA Arcade are here and now, MIST is already in its 3rd incarnation, and why would it be the final edition? As these boards gets more popular, people experiment more, demands increase etc.

I am happy to see enhanced versions but I wish they would change the name slightly to distinguish like Mist v3 or Mist MKIII.

The nice thing is that FPGA gives people the freedom to choose, the individual user can chose to use TG68 or Apollo or whatever there will be on his FPGA system, be it an accelerator for a real Amiga, or a stand alone board.

I agree but don't ask about the time I suggested allowing the TG68 to work on the Apollo accelerator for "compatibility" purposes. I received quite a response and I'm not sure if Gunnar was joking or peeved (maybe some of both). Anyway, the TG68 has been designed with compatibility as the main goal while the Apollo core has been designed with performance as the main goal. Users can have the best of both worlds with an FPGA as it is programmable.

Relevant how? The operating system is irrelevant, the software is irrelevant, the relevant games and applications are already reimplemented or ported or even emulated on relevant systems.

Relevant as in a healthy alternative platform which people can enjoy instead of declining to extinction.

I am not pessimist, I know that it will all go well. I am merely eyerolling over the delusionism in the community, it is not being negative, it is being realistic.

I am in this for the fun and entertainment :laughing:

Sarcasm! Now that is a good way to handle the situation .