Poster | Thread |
pixie
 |  |
Atari Falcon custom chips decap initiative Posted on 26-Jun-2025 20:10:27
| | [ #1 ] |
|
|
 |
Elite Member  |
Joined: 10-Mar-2003 Posts: 3474
From: Figueira da Foz - Portugal | | |
|
| |
Status: Offline |
|
|
BigD
|  |
Re: Atari Falcon custom chips decap initiative Posted on 26-Jun-2025 21:54:28
| | [ #2 ] |
|
|
 |
Elite Member  |
Joined: 11-Aug-2005 Posts: 7551
From: UK | | |
|
| @pixie
I thought the main benefit was an off the shelf DSP! The A3000+ had one too! I don't see how this benefits anyone? _________________ "Art challenges technology. Technology inspires the art." John Lasseter, Co-Founder of Pixar Animation Studios |
|
Status: Offline |
|
|
matthey
|  |
Re: Atari Falcon custom chips decap initiative Posted on 26-Jun-2025 23:08:04
| | [ #3 ] |
|
|
 |
Elite Member  |
Joined: 14-Mar-2007 Posts: 2734
From: Kansas | | |
|
| pixie Quote:
As seen on the FB page the falcon chips were decapsulated and are ready for reverse engineering. Would a similar solution benefit Amiga, for winuae, FPGA... ?
|
I expect there is little to be gained from decapping the Amiga chipset at this point and I would not be surprised if some of the custom chips have already been decapped. Most Amiga chipset cores have been reverse engineered and mostly debugged. There are schematics floating around for some of the chips (Alice chip schematic is at the following link).
https://github.com/nonarkitten/amiga_replacement_project
Jeri Ellsworth has the physical schematics for some of the chips as well. Schematics are more useful and less time consuming than decapping chips to find the logic. Decapping is more useful if trying to recreate the chip as close as possible but new ASIC custom chips are unlikely. Schematics and decapping are more likely to help with debugging and verification at this point. The most useful may be the chipset Verilog code including for unreleased chipsets like AA+ and AAA. We do not know that these are missing. Michele Battilana may have them.
The largest problems are the 68k CPU chips. The 68000 was successfully decapped which aids in reverse engineering the microcode and nanocode and allowing a cycle exact core which uses the original microcode and nanocode.
I'm officially building a microcode-level 68000 core https://gendev.spritesmind.net/forum/viewtopic.php?t=3023
The link above talks about the advantages and limitations of decapping for reverse engineering. It was very valuable for the 68000 which was relatively small but with complex and difficult to understand logic do to the microcode and nanocode. In comparison, the Amiga chipsets are relatively simple and straight forward as Jeri put it, more so than the C64 chipset. Higher end 68k CPUs increase in complexity and layers making them more difficult to decap, they no longer need to be cycle exact, and it would be more useful if the Verilog HDL code was available in order to work with as improvements can be made and bugs fixed without breaking compatibility. There are dozens of retro FPGA devices but they divide the market making it too expensive to license the 68k cores for their small market share. If someone were to realize the value of the 68k cores to retro IP and pay up to license the 68k cores and produce a retro 68k ASIC that could take most of the 68k FPGA market and expand it lower with the lower price, then all this reverse engineering development time could be avoided. The problem is the cheap shortcut guys want to replace the hardware and stick a retro label on a retro facade of a case rather than developing accurate and compatible retro hardware.
Last edited by matthey on 26-Jun-2025 at 11:12 PM.
|
|
Status: Offline |
|
|
kolla
|  |
Re: Atari Falcon custom chips decap initiative Posted on 26-Jun-2025 23:48:21
| | [ #4 ] |
|
|
 |
Elite Member  |
Joined: 20-Aug-2003 Posts: 3473
From: Trondheim, Norway | | |
|
| To “clone” the Falcon, one needs at least also a 68030 core (Suska was an attempt) and preferably also the 56k DSP (motorola 56001, opposed to the at&t 3210 used in the A3000+ and variants). So I’m curious where they will go with this, maybe replacement parts. _________________ B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC |
|
Status: Offline |
|
|
OneTimer1
|  |
Re: Atari Falcon custom chips decap initiative Posted on 27-Jun-2025 14:16:33
| | [ #5 ] |
|
|
 |
Super Member  |
Joined: 3-Aug-2015 Posts: 1245
From: Germany | | |
|
| Quote:
kolla wrote: ... curious where they will go with this, maybe replacement parts.
|
I don't know where they want to go, there are clones like the Mister that are compatible enough for most Atari and even Amiga software.
---
What could be done to get the Minimig core more compatible or effective would be a bounty system like it existed for Aros.
You know, people paying money to get a specific error removed or feature implemented.Last edited by OneTimer1 on 27-Jun-2025 at 02:50 PM. Last edited by OneTimer1 on 27-Jun-2025 at 02:22 PM.
|
|
Status: Offline |
|
|
pixie
 |  |
Re: Atari Falcon custom chips decap initiative Posted on 27-Jun-2025 17:28:37
| | [ #6 ] |
|
|
 |
Elite Member  |
Joined: 10-Mar-2003 Posts: 3474
From: Figueira da Foz - Portugal | | |
|
| @BigD
Quote:
I thought the main benefit was an off the shelf DSP! The A3000+ had one too! I don't see how this benefits anyone? |
I wasn't thinking in anything of falcon for amiga, i was thinking in using the same aproach used here on falcon chipsets but on amiga to improve emulation, fpga._________________ Indigo 3D Lounge, my second home. The Illusion of Choice | Am*ga |
|
Status: Offline |
|
|