English Amiga Board

English Amiga Board (http://eab.abime.net/index.php)
-   News (http://eab.abime.net/forumdisplay.php?f=29)
-   -   News of Free 060 Like Apollo Core License (http://eab.abime.net/showthread.php?t=89374)

Djole501 07 November 2017 19:22

News of Free 060 Like Apollo Core License
 
Fast translate from: http://www.a1k.org/forum/showthread.php?t=63012

Hello,

Fast 68040 or 68060 CPUs are not to be found in quantities.
Therefore, it is difficult to build new 68060 cards.

The APOLLO team offers a number of very fast 68K maps -
but we think shared joy is a double joy.

We would therefore like to discuss the following suggestion:
To support hardware developers, we would like to make the following offer:

We offer hardware developers a
_FREE_ 68080 CORE that is
compatible with the 68040/68060
and is set to about 68040/68060 speed.

The core is FREE for hardware developers.
The core also contains the AMMX2 instruction set.
This means that software compatibility with vampire software such as new JPEG decoder, video player and some game ports remains.

The Core boots with over 100 Sysinfo MIPS and after some time down to 68040 @ 40/68060 @ 50 level.

The supported hardware cards are subject to some reasonable conditions. That means the FPGA has to be big enough to offer a complete, fully compatible core space. The cards must be intended for the general public. The offer is not valid for single pieces and Pershingraketen.


We are looking forward to your feedback.

Greetings
Gunnar

alexh 07 November 2017 19:39

This would be very interesting for Atari Falcon 030 accelerator projects. They are not going to get the full Apollo core any time soon.

It all depends on compatibility.

[Edit: Just read the small print. No MMU and no FPU so useless for Atari Falcon 030]

grond 08 November 2017 11:24

Quote:

Originally Posted by alexh (Post 1197803)
[Edit: Just read the small print. No MMU and no FPU so useless for Atari Falcon 030]

There is a full FPU. A 68k-compatible MMU may come but cannot be promised as of now.

MartinW 08 November 2017 12:49

IMO Any initiative along these lines is to be applauded. I could sit here and say that 50Mhz is all well and good but most people want 100Mhz. Most people want MMU. Etc., etc.

But the bottom line is that there are a finite amount of real 68060 processors out there. And we're all merrily overclocking the hell out of them. So I for one would be most grateful if the day ever comes that mine dies for the chance to have a modern drop in replacement.

Sure, right this second there are likely some limitations (I don't know the detail, nor do I really care too much as an end user). But on the face of it, the general direction of this sounds great.

S0ulA55a551n 08 November 2017 15:36

This same post is a 4 page argument over on Amiga.org lol

Locutus 08 November 2017 16:27

The actual announce thread on A1k is pretty....something, lets put it like that.

Something is being offered, but not answers to reasonable questions :-/

Akira 08 November 2017 16:45

I think this is cool, it could help some people out when they have an idea but not the grunt to do a full 680x0 compatible core.
Despite how you feel about this particular 680x0 implementation, one can't deny it was a lot of work and that offering it to hardware developers free of licensing charges is very commendable.

However what I would REALLY want to see is a "competing" 680x0 core coming up from another team. It could only make some healthy competition with benefits for all. There's also the fact that this core will fit only a certain type of FPGA, so more options would be great. Like, what if you *just* want the 680x0, and not the aMMX or whatever else added to it?

Anyway, a good move, I believe :great

alexh 08 November 2017 19:02

From what I can tell the small print states that this offer is only open to companies building large batches of boards.

It is not available in low-quantities (or one off) for home-brew designs.

alexh 08 November 2017 19:05

Quote:

Originally Posted by MartinW (Post 1197960)
I for one would be most grateful if the day ever comes that mine dies for the chance to have a modern drop in replacement.

Unlikely anyone will do an FPGA->PGA adapter to fit to existing accelerators. It is for new cards.

sean_skroht 08 November 2017 21:37

Someone could create an accelerator to compete against the new non-existant Blizzard cards with these cores. Why hell, if that Blizzard guy ran out of real 68060 CPU’s he could use the Apollo cores. :crazy

grond 09 November 2017 11:39

Quote:

Originally Posted by alexh (Post 1198034)
From what I can tell the small print states that this offer is only open to companies building large batches of boards.

It is not available in low-quantities (or one off) for home-brew designs.

No, Gunnar explicitly stated on a1k that this offer is also open for hobbyists. However, he now commented that he is in contact with two interested parties which means that his support capacities for this type of project are likely consumed for the near future.

grond 09 November 2017 11:46

Quote:

Originally Posted by Akira (Post 1198009)
However what I would REALLY want to see is a "competing" 680x0 core coming up from another team. It could only make some healthy competition with benefits for all. There's also the fact that this core will fit only a certain type of FPGA, so more options would be great. Like, what if you *just* want the 680x0, and not the aMMX or whatever else added to it?

Well, anyone would appreciate options but to be realistic: the 080 is incredibly complex and the likelihood of anybody else doing anything that comes even remotely close to it is zero. I'm a microchip developer by profession and I have seen the VHDL code. It is too complex for me to even understand more than the rough structure or tiny isolated fragments of it. This is also why I laugh at all demands for open-sourcing the code. There just isn't anybody who could pick it up. It is as it is: a competitive CPU isn't something that some hobbyists can develop. We should be grateful that some equally brilliant and persistent professional CPU developers picked the Amiga and 68k CPUs for a hobby.

daxb 09 November 2017 12:55

That means that all stands and falls with Gunnar. Not a good basis, right?

grond 09 November 2017 13:12

Quote:

Originally Posted by daxb (Post 1198222)
That means that all stands and falls with Gunnar. Not a good basis, right?

Well, he is more alive than Motorola and Commodore... Apart from that he is not the only one with a copy of the code, there are at least four others that have access to the repository.

Overflow 09 November 2017 13:47

Gravity outlived Newton.

grond 09 November 2017 13:53

Quote:

Originally Posted by Overflow (Post 1198228)
Gravity outlived Newton.

Only because Einstein filled the gap... :agree

fpgaarcade 09 November 2017 15:13

Quote:

Originally Posted by grond (Post 1198217)
Well, anyone would appreciate options but to be realistic: the 080 is incredibly complex and the likelihood of anybody else doing anything that comes even remotely close to it is zero. I'm a microchip developer by profession and I have seen the VHDL code. It is too complex for me to even understand more than the rough structure or tiny isolated fragments of it. This is also why I laugh at all demands for open-sourcing the code. There just isn't anybody who could pick it up. It is as it is: a competitive CPU isn't something that some hobbyists can develop. We should be grateful that some equally brilliant and persistent professional CPU developers picked the Amiga and 68k CPUs for a hobby.

Grond, with all respect that statement is complete b@llocks.
As a professional CPU designer, if somebody who is experienced cannot read, understand and modify the code, then it's not as good as you think it is.

Open source designs such as :
https://en.wikipedia.org/wiki/RISC-V
https://github.com/freechipsproject/rocket-chip
https://github.com/SpinalHDL/VexRiscv

This is a decent modern, pipelined CPU core which is easy to read.
Yes, it's RISC not CISC as you are probably about to argue, but it doesn't make That much difference.
223MHz in Artix-7 with MMU, that's over 250 DMips.

/MikeJ

fpgaarcade 09 November 2017 15:18

Although I will give you contortions in the TG68K are quite challenging. I'm having to use our ASIC formal equivalence tools to unwind it..

grond 09 November 2017 15:20

Quote:

Originally Posted by fpgaarcade (Post 1198253)
This is a decent modern, pipelined CPU core which is easy to read.
Yes, it's RISC not CISC as you are probably about to argue, but it doesn't make That much difference.

It makes a world of a difference and you know it.

fpgaarcade 09 November 2017 15:23

The instruction decoder is a massive pain, but the instruction set is highly ordered (it was originally PLA decoded) and the same tricks are possible now. It's only a 16 bit opcode for goodness sake!


All times are GMT +2. The time now is 09:46.

Powered by vBulletin® Version 3.8.8 Beta 1
Copyright ©2000 - 2017, vBulletin Solutions, Inc.

Page generated in 0.07622 seconds with 9 queries