17 March 2021, 03:38 | #41 | |
Registered User
Join Date: Sep 2006
Location: Thunder Bay, Canada
Posts: 4,323
|
Quote:
|
|
17 March 2021, 03:42 | #42 |
Registered User
Join Date: Dec 2015
Location: USA
Posts: 2,912
|
No wai, I'll just use my Ifone to watch people using Amigas instead, thanks for nuffin!
|
19 March 2021, 04:50 | #43 |
Banned
Join Date: Dec 2018
Location: Australia
Age: 51
Posts: 99
|
CPLD or FPGA reproduction of the original chips does not mean emulation.
Make each chip separately and install a module. I've got an FPGA Agnus / Alice for A500 which I'm still working on. Once we have reproduction, we can work on improvements. The programmable ligic is where I agree with Sukko Perra that they should not be owned by anyone like SAGA. They will need to be an open community project. But also perhaps free of GPL infection, they ought be MIT licensed? The more you look at the Amiga designs, the more you see its like leggo blocks creatively put together in different ways. Gary, Gayle, Bridgette, Paula and others can be recreated in CPLD. What wont fit in a CPLD can fit in an FPGA instead. So Amiga with maybe 5 FPGA chips inside it? |
19 March 2021, 11:16 | #44 | |
Registered User
Join Date: Feb 2015
Location: Sweden
Age: 50
Posts: 2,953
|
Quote:
the day *will* come when we run out of NOS replacement chips or even working used ones so a CPLD/FPGA drop in replacement for each one of them is probably something that needs to happen at some point in the future.. And no one would want such projects to be closed. On this I think we have agreement across the community.. The Apollo team is pursuing their own dream of an enhanced Amiga and it makes sense they build this dream around AROS which they can control themselves.. But the more time passes the more likely they will eventually be considered as (yet) another separate platform... |
|
19 March 2021, 11:51 | #45 | |
Registered User
Join Date: Sep 2009
Location: Norway
Posts: 1,711
|
Quote:
I agree with Thomas Richter here. |
|
19 March 2021, 15:25 | #46 |
Guru Meditating
Join Date: Jun 2014
Location: England
Posts: 2,339
|
Yup, Thor has absolutely hit the nail on the head here. Anyone talking about a new Amiga as becoming another platform is delusional.
|
20 March 2021, 00:00 | #47 | |
Registered User
Join Date: Feb 2015
Location: Sweden
Age: 50
Posts: 2,953
|
Quote:
For the rest of us, it will be about accepting that we had back then was a very good solution to a problem that existed in the mid 80s. For 2020's problems we have new solutions.. But we still like the nostalgia factor. Anyhoo.. |
|
21 March 2021, 10:00 | #48 | |||
Registered User
Join Date: Mar 2018
Location: Hastings, New Zealand
Posts: 2,584
|
Quote:
Quote:
Quote:
|
|||
21 March 2021, 18:27 | #49 |
Registered User
Join Date: Feb 2015
Location: Sweden
Age: 50
Posts: 2,953
|
Who here thinks it would be a good idea to add a STM32H7 to a new Amiga motherboard to be used like i n the CS labs accelerators.. (Although how it was implemented ios not open source/common knowledge)
The chip itself is something like $-10 and readily available. Pros, Cons? |
21 March 2021, 20:11 | #50 | |
Registered User
Join Date: Aug 2014
Location: Netherlands
Posts: 697
|
Quote:
It feels a bit "wrong" to have a support controller that is more powerful than the computer it is supporting. |
|
21 March 2021, 23:09 | #51 | |
Registered User
Join Date: Feb 2015
Location: Sweden
Age: 50
Posts: 2,953
|
Quote:
It also has a built in jpeg decoder which I'm sure could be utilized. And there's also a pretty powerful FPU on that thing which could probably be used for something.. But... if we ignore all other stuff it can do and only focus on the RTG part where you can have a $7 chip on the motherboard that can give you Full HD RTG. Third party graphics card are almost as old as the Amiga itself. So not blasphemy in the same way as it if the 480MHz CPU would be used to run software in the way Blizzard PPC ran stuff alongside the 68k cpu... |
|
22 March 2021, 02:38 | #52 |
WinUAE 4000/40, V4SA
Join Date: Apr 2020
Location: East of Oshawa
Posts: 538
|
I'd simply be happy to have a new source of 68k family CPUs and the set of custom chips. Even better, open-use VHDL or Verilog sources for the above, for FPGA synthesis. As long as we have that, people can come up with whatever Amiga systems they want whether it's discrete chips on a board (classic systems and remakes of them), systems on a chip (Vampire), or somewhere in between. If you want, you could even create your own updated versions of the chipsets this way.
Has anyone actually focused on this, just describing the existing chips in VHDL or Verilog, and putting those files out there? Apollo Team must have something like that for their Vampire FPGA implementation but of course that's closed. |
22 March 2021, 08:07 | #53 | |
Registered User
Join Date: Feb 2015
Location: Sweden
Age: 50
Posts: 2,953
|
Quote:
http://www.bambi-amiga.co.uk/amigahistory/clone-a.html Of course, you also have the FPGA Arcade, MIST and variants.. etc.. So yea the idea is not new.. And if you search GitHUb etc you'll probably find a few more.. But It's already been mentioned here that what people probably want is the entire chipset done as a drop in replacement (chip by chip). However, I feel that discussion is for another thread. The focus of this thread is to discuss a new Amiga motherboard that would hit the optimal balance with most power for your cash. Let's focus on that and the Chipset as FPGA /CPLD replacements discussion can be its own thing ;-) |
|
22 March 2021, 10:39 | #54 | |
Banned
Join Date: Dec 2018
Location: Australia
Age: 51
Posts: 99
|
Quote:
Firstly it could emulate the original CIA chips. Damien Stewart in Melbourne has already made a PPC emulation of CIA's for Amiga OS4 I've been discussing with him to port over to the STM32 for software which absolutely requires the 8520 CIA chips. PERIPHERAL Gotek flash floppy USB Ethernet DMA controller for IDE SPI, I2C, I2S, Just about every comms standard in existance... If used to provide FPU, then we could make use of those 75MHz 68EC060's. CPU Coprocessor FPU DSP JPEG codec decode MPEG key frames accelerating playback Encryption, Compression These tasks can be limited to libraries so as to prevent further Amiga fragmentation by some applications being written in ARM instead of 68K. The secuity environment of the STM32 can prevent other ARM code execution preventing security violations. The CON of the STM32 is risk of fragmentation of Amiga like PPC WARP. So disabling the ability to execute ARM code outside of OS libraries is perhaps a wise precaution. I already decided to use this chip (or STM32MP157) before i had heard about the Warp cards. |
|
22 March 2021, 12:13 | #55 | |
Registered User
Join Date: Sep 2013
Location: Beeston, Nottinghamshire, UK
Posts: 238
|
Quote:
|
|
22 March 2021, 14:39 | #56 |
WinUAE 4000/40, V4SA
Join Date: Apr 2020
Location: East of Oshawa
Posts: 538
|
Amiga Replacement Project seems to be along the lines of what I'm thinking, yeah. Although based on the readme, this "AGA-on-ECS" thing instead of straight-up AGA implies there's no plan to actually implement Alice and Lisa, only try to cram some of their functionality into the ECS Agnes and Denise chips instead. Also, no activity in nearly half a year is generally a bad sign.
None of these address reproduction of 680x0 either. While I'm sure there are easy-to-find 68000 VHDL or Verilog cores, what about the '40 or '60? Anyway, enough derailing from me. I don't like the idea of having modern ARM cores in an Amiga system that isn't NG/TG, although I could accept it if they're only used by board firmware (i.e. not even accessible to the OS). As PurpleMelbourne points out, it's a vector for bringing back the WARP civil war just with a different processor architecture. The idea of disabling ARM code execution "outside of OS libraries" sounds good, but given how the Amiga OS works would that even be possible? Seems to me that what the OS can do, any other application can also do. If the STM32 is only used like a PC southbridge for managing peripherals, network access, and non-Amiga buses, that would be okay. Just don't put it in a position where OS or app code can make direct use of it as a processor. |
22 March 2021, 23:39 | #57 | |
Registered User
Join Date: Feb 2015
Location: Sweden
Age: 50
Posts: 2,953
|
Quote:
The reason this particular chip is probably a good option in a vast ocean of other options, is the fact that its already out there on the Warp boards. So there should be some potential synergies. (Like using P96 for RTG and hopefully more) Shouldn't really be adding to the fragmentation. But yea, I agree - running software on the chip in some sort of hybrid solution is probably crossing a line.. OTOH.. Remember the old days, it was a completely normal thing for Amigans to expand their big box Amigas with non Commodore/Third Party hardware to improve the function of the machine.. so not exactly a new concept in Amiga land.. I think its mostly about people who are still into it to this day don't feel much nostalgia towards these third party solutions. And also a lot of people like to use "contemporary solutions". Meaning tech that was available back then. But that part is proving more and more difficult. (Like scouting the earlier mentioned DSP chip used in the AA3000+ project..) So while cool, a sensible balance need to be found or it becomes too niche within already a small niche. So at some point, we will need to make this "recreating the chipset(s) in FPGA as drop in replacements happen or start accepting some level of emulation/hacks using modern chips. The raspberry PI zero "hack" for Denise that brings HDMI out at a very small cost is such an example.. IMPO I think such hacks are ok. Make a lot of sense without ruining the authenticity of the Amiga experience. But a line was crossed for sure.. |
|
23 March 2021, 11:42 | #58 |
Registered User
Join Date: Jun 2018
Location: Stockholm
Age: 48
Posts: 265
|
Regarding recreating custom chips, this is what you need to do in order to reverse engineer them, http://www.experiment-s.de/en/gallery/ ,take them one-by-one, look at the pictures with the piggy-back adapters hanging in wires. This way you get access with an Oscope and LA as well as JTAG for reprogramming (because it is going to be a lot of that). This guy started doing it back in 2003 for the Atari custom chips. Very impressive. It takes a lot of time, patience and perseverance. A lot of trial and error time consuming testing with the replacement CPLD/FPGA in a real machine, many long nights and so on... It can be done but are you up for doing the actual work? Do you have the time to spare? Do you know the inner workings of the Amiga well enough? Is your verilog up to speed? Are you going to release your work for free into the public domain once you are done? I'm pretty sure nonarkitten will accept PRs (pull-requests) to her github repo should you have them coming. Make sure you agree with the non-commercial Creative Commons license first before any commit, and prepare to lawyer up to fight against the clones that will appear based on your work anyway. We know this happens in Amiga-land all the time.
Of course if you are designing a new Amiga it makes most economical sense to put all the custom chip behavior in one large FPGA together with some external RAM and the CPU (also in verilog/vhdl). This is basically what the V4 SA or the Minimig/MiSTer is. It might not be as fun though without Paula, Agnus, Denise etc doing the "chip-dance" on the actual dance-floor and no floppy drive eating disk in the background. Will the dance be as fun with many CPLDs/FPGAs on the floor replacing the now old ladies?. Idk. I guess it all comes down to how much of an Amiga purist you are. Perhaps even having the CPU in FPGA is a no no for some people, it needs to be genuine Motorola ASIC with a rev and a mask number on it?. Beauty is in the eye of the beholder. Single-board computers can feel a bit dull because of their limited expandability, however the RPi has become a very popular SBC and you can for example run emulation such as PiMiga on it (based on Amiberry) which is a very affordable standalone solution. I think before you decide whether a specific arm chip should go on your new board or not, you should focus on how to solve the scarcity problem of the custom chips, because I hope your plan for this new motherboard goes beyond pulling chips from old working Amiga motherboards. |
23 March 2021, 13:35 | #59 | |
Registered User
Join Date: Dec 2019
Location: Ur, Atlantis
Posts: 1,914
|
Quote:
Not saying one or the other is better, but it really can be an amusing to think about. |
|
23 March 2021, 14:20 | #60 |
Registered User
Join Date: Jun 2012
Location: San Francisco-ish, via Ireland
Posts: 114
|
The problem though, as others have mentioned, is that at some point we’ll simply have no choice. The old custom chips are going to eventually go away, and finding new (not everyone will agree that it’s necessarily improved) ways of preserving the platform is really the only way forward.
|
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
Thread Tools | |
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
SukkoPera's Open Source Hardware: A500/A600 Chip/Fast RAM Expansions | SukkoPera | Hardware mods | 208 | 05 February 2023 02:41 |
Open Video Toaster Install Guide, no hardware required | Pyromania | Amiga scene | 6 | 29 January 2021 13:44 |
Open Amiga Game Database - Open Beta | FrodeSolheim | support.FS-UAE | 3 | 01 May 2013 22:37 |
Pandora (Open Hardware) now available to pre-order | chiark | News | 474 | 27 December 2012 00:06 |
Amiga Software Index - The Big Bang - | sun68 | News | 5 | 14 December 2010 13:00 |
|
|