natami

[matthey]

Hi,
Personally I have no interest in licensing a closed source CPU core.

Hi Mike. First, I am not an official spokesman for the Apollo project. Gunnar has talked about releasing the Apollo core source to specific customers under a strict license (and for a hefty fee). You would have to contact the Apollo Team/Gunnar for details and they may not be set in stone yet. I understand that the Apollo core may not be suitable as your main 68k core for various reasons. Perhaps you could optionally allow your customers to use the Apollo core in the fpgaArcade?

The T68K is continuing to be improved and has reached a good level of compatibility with a 68020 - it's not perfect yet. Next generation cores (I am writing one) are table based and will be much more efficient, especially as we are learning from the 68000 die scans how the microcode is laid out in the original.

It should be possible to get >100MHz 68020 performance in the cheap FPGA the replay
uses.

Full 68020 compatibility is not easy while 68000 compatibility is fairly easy. The Apollo core will not be fully 68020 compatible but very 68000 compatible. While most of the 68020 changes were good, we feel that some were mistakes. The Apollo will likely drop and trap some rarely used instructions and addressing modes. That may not be appropriate for the main fpgaArcade 68k core. Good luck in full 68020 compatibility for the T68k .

The Apollo core will have some go fast features that will take a lot of work to duplicate like pipelining, Superscaler, instruction combining, new powerful instructions and addressing modes, sophisticated instruction and data caches (including snooping), loop optimizations, branch cache and prediction, etc. How fast the core clocks doesn't tell the whole story but 68020+ support at over 100MHz does look probable. It will be interesting to compare the performance of the TG68k in the fpgaArcade to the Apollo core in the Natami.

[Darrin]

It should be possible to get >100MHz 68020 performance in the cheap FPGA the replay
uses.

/MikeJ

That should be interesting.  For years my main Amiga used an overclocked 68020 at 28MHz.  Now I use a couple of 68040 CPUs at 40MHz.

How do you expect a soft 68020 at 100MHz to stack up against the real 68060 at 100MHz on the daughterboard?

[Johan Samuelsson]

Oooh! Do I hear the sounds of a healthy competition?
Go on teams! Fight! I will buy both machines!

[Thorham]

How do you expect a soft 68020 at 100MHz to stack up against the real 68060 at 100MHz on the daughterboard?

If the soft 68020 has the same internal timings as a hard 68020/68030, then it will only be twice as fast as a 50 mhz 68030. In other words, the '060 will probably be around four to five times faster still.

[Darrin]

@ Thorham:

Thank you for that info.  I've never owned a 68060 board so I've got no practical experience to gauge the speed difference of one.  

[motrucker]

I belive they said boards will be in 500-700 eur, so kind of SAM range.

That lets me out. That kind of pricing will severely limit their sales potential across the board.

[matthey]

If the soft 68020 has the same internal timings as a hard 68020/68030, then it will only be twice as fast as a 50 mhz 68030. In other words, the '060 will probably be around four to five times faster still.

The timings should be significantly better than the 68020/68030 for both soft cores. Most instructions in the Apollo core will have similar or better timing than the 68060. Memory speed is faster and caches are bigger. There is a link stack for subroutines which the 68060 does not have. The Apollo core should be able to outperform a 50MHz 68060 at least. We'll have to see compared to a 68060@100MHz .

[ToddH]

Starting to have my doubts after seeing this posted over at the Natami site:

Dear fans of Natami!

Sorry for the long silence. After discussing important matters on the team, we hereby announce the need to reorganize the project.

Due to unreasonably high expectations on the team side regarding time frames and foremost forgetting about Thomas' original intentions for what Natami is supposed to be, Thomas has decided to seclude himself as of now and to peacefully finish what he has begun, without having to follow unrealistic time frames and expectations.

We as a team feel that all you strong Natami followers deserve a better treatment than being confronted with a wall of silence, while eagerly awaiting news and progress.

Natami is going to be finished, yet not necessarily under that exact
name (for reasons not to be debated). What name it will be, is Thomas' choice.

So, if you don't mind a Natami-like machine being officially available at some (as of now yet uncertain) point in time and want to stick with it anyways, you're very welcome. Should you think that other projects seem more wortwhile in the light of
things, please follow your wishes.

That kind of fairness and openness is the least we have to treat our strong Natami fans with.

Best regards to all you great supporters
Team Natami

[jj]

Whilst I think this project could have been cool.  I think this turn of events is not exactly surprissing.  I am guessing several walls have been hit

[Kesa]

Whilst I think this project could have been cool.  I think this turn of events is not exactly surprissing.  I am guessing several walls have been hit

Care to speculate? Anyway i think this is bad news. I was waiting for one of these. What's wrong with the name?

[-BobW-]

Care to speculate? Anyway i think this is bad news. I was waiting for one of these. What's wrong with the name?

I think this is probably good news.  The scope of the project had gotten way out of control.  It sounds like Thomas has now refocused on a more realistic scope.  The Natami that everybody wanted would have taken years to complete.

I wish them luck.

[Darrin]

Care to speculate? Anyway i think this is bad news. I was waiting for one of these.

The Natami specs just kept getting better and better, plus the design of a new and improved CPU must count as an entire project on its own.  There was a hell of a lot of work going on here and (in my opinion) they might have been better off producing several motherboards, each one improving on the next, or perhaps taking a similar route to the FPGA Arcade where you can sell a functioning "super A1200+" board and then work on an expansion board to add additional features and turn it into an Amiga 5000.

This is a real shame, but at least the project is still alive.  Meanwhile we still have FPGA solutions for classic fans while we wait.

What's wrong with the name?

Perhaps AmiNat sounds better.

[wawrzon]

funny part is that the guy behind the project leaves the team, so who remains and what will they do without him, maintain the forum?  
however it looked dull since longer time, i really wondered how a private person is able to maintain the project single handedly year for year. this had to happen.

good thing is the fpga arcade remains for those who need it even if it exhibits a little similar pattern.

[TheBilgeRat]

Ah well, the naysayers called it on the nose.  So much for the Natami.

[commodorejohn]

...and feature creep claims yet another victim

I'm still hopeful that something will come of this, but I'm not going to be holding my breath...

[JimS]

...and feature creep claims yet another victim

I'm still hopeful that something will come of this, but I'm not going to be holding my breath...

Babbage Syndrome. Get almost done, then think of something even cooler.