@CodeSmith

I'm not so sure. Multicore PPC well into the GigaHertz domain can be had below 40USD (in volume). You're not going to get that sort of performance from an FPGA in the same price range.

But from an electrical engineers perspective, there's very little difference.

Any design I can imagine which has a PPC to solve an Amiga related problem could be redone with an FPGA. For instance the "rip out laptop motherboard, replace with T1020+SB850+MXM design" from another thread. Just as easily done with an FPGA, except the "how do you get PCIe to both GPU and SB" question which is sort of tricky in an FPGA (PCIe hardware typically only supports 1 link).

But for this to be useful as an expander of user base, it would need to ADD something. Like new coprocessors not seen before.

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.

@CodeSmith

Oooh. Just remembered the first PPC Macs. 64 bit data bus on processor, 32 bit system bus from 68040 design. Won't run Linux, cause it's just too weirdly wired.

In fact, I have (had? not sure what I kept) (access to? don't remember even whether I was just borrowing one of them) 3 different Macs, all on the "ten worst macs ever designed" list. Finding an old mac that could run ANY sort of Linux was a bit of an adventure. Custom chipset does not equal awesome

_________________
This weeks pet peeve:
Using "voltage" instead of "potential", which leads to inventing new words like "amperage" instead of "current" (I, measured in A) or possible "charge" (amperehours, Ah or Coulomb, C). Sometimes I don't even know what people mean.

@olegil

I'm not thinking about it from the point of view of "making the amiga relevant again", rather from the point of view of a cool 1980s style computer - the sort of thing a smart teenager could understand in its entirety (try writing a game these days with full graphics and sound, that boots from USB and doesn't use anything other than UEFI services and bitbanging I/O ports). From that angle I'm not concerned about PCI buses or high speed memory controllers. The sort of computer I'm thinking of would use either SRAM or the simpler kinds of SDRAM and the biggest expansion concerns would be over making sure that it can support the newer type of displays (HDMI, DP, etc) and desktop buses (basically the latest flavour of USB). Both of those can be done by FPGAs, and if you're feeling a bit lazy the latter could be done by one of the newer Atmel chips with builtin USB support, talking to the FPGA over a fast three wire serial bus.

Last edited by CodeSmith on 25-Sep-2015 at 09:39 AM.
Last edited by CodeSmith on 25-Sep-2015 at 09:38 AM.

@CodeSmith
Quote:

I'm quite convinced that the future of the amiga is in virtualized 68K in FPGAs by similar projects to this and AROS or some other OSS for the OS.

yes. im convinced the same, even if the insistence on ppc architecture as an "official" consequence and trying to push it through no matter what might destroy the scene at some point beyond repair, as it has been inflicting lasting if likely not yet critical damage in the past. luckily it seems the majority of amiga enthusiast have steered clear of os4 ppc forums and the like, as they have steered clear of the insane ppc expansions at its time.

@wawa

If you lose current PowerPC users and developers, your vision of 68k future will never become reality. Most remaining skilled software developers already made their choice... and it was not 68k. Question you should answer first is: how to get them back.

@pavlor

Quote:

If you lose current PowerPC users and developers, your vision of 68k future will never become reality. Most remaining skilled software developers already made their choice... and it was not 68k. Question you should answer first is: how to get them back.

you are talking about morphos users and developers, i take it. observing polish forums, where a big part of them are active, there is some interest towards fpga softcores even among them. be it really the case or not, it will occur as soon as the solution is widely available, i wouldny like to impose anything on the community

Last edited by wawa on 26-Sep-2015 at 02:04 PM.

@pavlor


by offering a bigger market and better development tools

@cdimauro

Quote:

cdimauro wrote: @pavlor Quote: pavlor wrote: @cdimauro 4 months means old? You live fast life. More or less. I think that software evolves much much faster than hardware.

Regarding this statement, from WinUAE 3.2 Beta 14: 64-bit JIT fixes. JIT FPU also fixed.


@itix

Quote:

itix wrote: @cdimauro In theory they would, but at that time this kind of software was already optimized for real FPU.

But not for my plain 68(EC)020, which was much more common that an Amiga system with an FPU.


@NutsAboutAmiga

Quote:

NutsAboutAmiga wrote: @cdimauro It's kind of slow to have to jump into a library to do FPU instruction. It most likely makes more sense to have two versions of same binary, with FPU and without FPU.

Yes, but the problem here is different. If you have a machine without an FPU, you have to emulate its features, and a 68020 version of such code is much, much faster than a plain 68000 one.


@kolla

Quote:

kolla wrote: @cdimauro Quote: cdimauro wrote: @itix: at least the mathieee* libraries would have greatly benefited of 68020-specific versions. But they don't need to be in kickstart. There are third party math libraries, optimized for all the CPUs with and without FPUs, you can knock yourself out already.

Even for a 68020 without FPU? Interesting. I wasn't aware of it. Have you some link to such stuff?

Anyway, it would have been interesting to measure differences between ClassAct/68000 and Reaction/68020, since they aren't number crunching components, but "system" ones which have to deal with lists/nodes, "method" invocation, etc.. If there's some impact with them, it's likely that a 68020 can have impact even on the Kickstart.


@OneTimer1

Quote:

OneTimer1 wrote: @cdimauro Quote: cdimauro wrote: It's good for 68000 users, but I'm greatly disappointed to hear that, because it means that the Kickstart 3.x wasn't optimized for 68020+ processors. IMO at the time a 68020 should have been a minimum requirement, ... I don't know if AOS3.x for A1200/A4000 has 68020 optimisation or not,

There are a few of them.
Quote:

but many software for AGA was never made for 68000, you will find some software that was never compiled for 68000 because the Authors thought a 68000 CPU with 8MHz RAM will never be enough for this software. Therefore 68020 command set on AGA is a must.

Absolutely.
Quote:

But when it comes to performance the FPGA-68k is just some kind of 68000 on speed, a 68k variant with more commands per clock as the original 68k and with clock rates never available on a naked 68000.

You can get a good 68K design with an FPGA with several resources. FPGAs nowadays are available also with high clock speeds (compared to the past; but they are much slower than a regular CPU, of course).

But there are two problems here: the cost of a very good FPGA, which is too high, and the time to develop something cool with it, which is very long.


@blizz1220

Quote:

blizz1220 wrote: BTW I think that it should be 300 Mhz Coldfire not 30 ??? What would be the point ?

Coldfires aren't 100% 68000 or 68020 compatible, so their use for Amiga-like projects is out of question.


@kolla

Quote:

kolla wrote: @OneTimer1 Quote: But when it comes to performance the FPGA-68k is just some kind of 68000 on speed, a 68k variant with more commands per clock as the original 68k and with clock rates never available on a naked 68000. The 68SEC000 in the Minimig is not on FPGA, it is a real physical CPU, and it is fast enough too. The 68000 has developed in newer and improved version all the time, just rarely used in amiga products.

If the 68SEC000 is not solder (but according to olegil, it seems so), in theory you can replace it with a small FPGA board which implements a much faster 68020+ CPU, using the same 68000 socket (and signals).
Quote:

Anyways - having the essential OS3.9 components rebuilt as 68000 compatible should not be a big issue. I am puzzled by how workbench.library became 020+ on rev 45.130 (I use 45.129 on my Minimig), without any word in the docs/changelog about why.

Can't you use an application which intercepts 68020 instructions, emulating them on a 68000 CPU?


@number6

Quote:

number6 wrote: @thread Just repeating these links in case it might answer to anything concerning plans or speak to the issue of who would communicate about what: ACube and A-EON announce One Vision on future Amiga co-operation

14-Dec-2013: "We are pleased to be working with like-minded Amiga enthusiasts and together we will work to support and expand the Amiga user base"
Quote:

and A-EON Technology and ACube Systems strengthen strategic alliance

21st May 2015 (after one year and half): the two companies agreed a joint development plan to ensure the future of Classic and Next-Generation Amiga hardware and software development
[...]
"During our meetings we had very good MiniMig Plus prototype ideas about near future hardware and software development, really can't wait to realize all of them."
[...]
"The combined technical and financial resources of A-EON & ACube should help secure future success."
[...]
"I'm more optimistic now about the future Amiga scene than I have been for a very long time."

@cdimauro

So you suggest a 68020 "emulation layer" for 68000? Is that also you would want to run 060 software on 68020?

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@cdimauro

Not sure what you're trying to say with the quotes from my links above.

But the Acube connection with A-Eon was at least partially detailed regarding X5000:

Quote:

Once the system is commercially available, Max "mx3" Tretene of ACube will work on future firmware improvements and updates. Another sign of our close cooperation with that company.

Source

#6

Last edited by number6 on 03-Oct-2015 at 01:51 PM.

_________________
This posting, in its entirety, represents solely the perspective of the author.
*Secrecy has served us so well*

@kolla

Quote:

kolla wrote: @cdimauro So you suggest a 68020 "emulation layer" for 68000? Is that also you would want to run 060 software on 68020?

Exactly. Of course, it'll be much slower, but it depends on how many 68020+ instructions will be executed.

However, I don't think that the newest workbench.library's code is executed so often.

@number6

Quote:

number6 wrote: @cdimauro Not sure what you're trying to say with the quotes from my links above. But the Acube connection with A-Eon was at least partially detailed regarding X5000: [...] #6

I was referring to this topic: the Minimig+.

@Thread

One year has passed since I opened this thread, ACube remained silent about the Minimig+.

I don't know if there was any production run of a Sam460 during the last 12 month but I have not seen any new Sam460 user here.

To me it seems as ACube has finally left the Amiga Scene.

The Minimig+ might become obsolete soon, when the standalone version of Gunnar von Boehn's FPGA-Amiga becomes reality.

Aeon will produce the next low cost version of AOS4 hardware without ACube and users stopped asking for ACube months ago.

@OneTimer1

Max from ACube has been very active recently in developing the new Ringhio Notifications that is now part of the Enhancer Software. Without his help, we would not have the updated including the new Progress Bar support.

Enrico from ACube has helped logistically with the boxed CD edition of the Enhancer Software, so we are very grateful to ACube for their help.

_________________
Amiga Kit Amiga Store
Links: www.amigakit.com | New Products | A600GS

@OneTimer1

Quote:

The Minimig+ might become obsolete soon, when the standalone version of Gunnar von Boehn's FPGA-Amiga becomes reality.

it was obsolete to start with. good idea to shelve it and forget. if anything apollo/vampire is the way to go. not sure why to complain in this situation, where solution is at hand.

finaly has two more weeks become two more years !!!!

@wawa

Quote:

wawa wrote: it was obsolete to start with.

There was no Apollo core when they started and all (free) 68k cores reached less speed than a 68030 / 25MHz - so it looked logic. Schematics for the Minimig or simple 68030 accelerator card where open source therefore development should have been easy.

If they had started early enough, they could have sold Minimig+ during the last two years, instead they ran into the Osborne-Effect ( https://en.wikipedia.org/wiki/Osborne_effect ) without need.

But with the existence of Apollo core, the original design of the Minimig+ became obsolete, at least if the Apollo Minimig (or however it will be called) is sold for less than 500€.

Last edited by OneTimer1 on 17-Sep-2016 at 12:20 PM.

@wawa

Quote:

you are talking about morphos users and developers...

You know, there are skilled OS4 developers.
Yes, MorphOS has attracted some of the best developers that were still working at the end of the Amiga's production life.
But that is partly because some of them contributed to it.

OS4, at least at first, had a smaller developer base.

_________________
Amiga! "Our appeal has become more selective"

So what is going on with Acube these days?

_________________

@AmigaMac

This is aCube Project

http://amigaworld.net/modules/newbb/viewtopic.php?mode=viewtopic&topic_id=41191&forum=33&start=80&viewmode=flat&order=0

@OneTimer1

Quote:

There was no Apollo core when they started

im not sure, i think there was, maybe you were not aware of it. at least vampire itself was underway afair and fpgaarcade and mist or whatever platform able to run tg68 have already delivered the performance in the range we are talking about.

this project was simply meant to catch up, but hard to say if it ever was more than a concept.