@Petah

FPGAs are expensive and slow.

You can get PPC SoC with a clock frequency of 400MHz for a few bugs but an FPGA and big enough for an equal performing Softcore will cost a fortune.

This might be different with abandoned CPUs like the 68k. A 200MHz 68k (with full 32Bit, big caches, FPU) would be a sensation (for us), the same with PPC would be a flop.

Last edited by OneTimer1 on 07-Dec-2019 at 02:14 PM.

Voted for 3.

Powerpc socs / cpu’s are much faster and cheaper then fpga.

Even Phase 5 allready had a CyberstormPPC G4 on their tables.

Voted for 3.

It would be slow and expensive.

@Petah

I voted for the third option for the same cost and performance reasons that others have stated in this thread.

If I was spending a lot of money for new PPC hardware to put in my (insert favourite Amiga) then I'd want something featuring a P or T series SOC so that performance and video options would be on par with current AmigaOne systems.

PPC accelerator around p1022...
1) bc Trevor has plenty of those
2) core1 could emulate 68k and core2 could emulate 603 (BlizzardPPC mode)
3) alternatively core1 could emulate HombreCPU and core2 could handle HombreGPU
4) diu could be used as 24bit framebuffer output
5) alternatively diu could do 3D (for AOS3 and/or for AOS4) with the help of the core2

Yeah. Silly stuff.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@Kronos

Quote:

Kronos wrote: @Deniil715 Adding an (FPGA) AGA to a NG system makes about as much sense as adding a FPGA PPC to an classic Amiga. None at all if you wonder.... FPGA+NG: What SW needs the chipset but still works in an NG OS (running on Phase5-HW or UAE)? Little to none I'll assume.

That's a very good point. Would never work out of the box. Then build a complete Amiga on PCI, but then again, we have MiniMig etc.

The whole thing gets pointless.

It's all so contradictive - People want a new Amiga, better and faster. But then they want the Classic HW with hopeless gfx design and 8-bit sound. We present new HW with fast CPUs, but no we want Classic. But then they want upgrades to the Classic but don't get that upgrade to Classic is the same as NG.

_________________
- Don't get fooled by my avatar, I'm not like that (anymore, mostly... maybe only sometimes)
> Amiga Classic and OS4 developer for OnyxSoft.

@Petah
'Twas no surprise to me when I voted for the third option to see that it is the one with the most votes.

As you've indicated, the development of FPGA is following the usual technology trend of increasing potency whilst decreasing in price. Right now, the class of FPGA required for a software implementation of even a single-core PowerPC (Power ISA) CPU is not economically viable, but I will answer this question as if you asked it 18 months from now.

The market for and FPGA implementation of PowerPC cores is extremely small. Outside of idle curiosity, or the nostalgic desire to run AmigaOS 4.x there is no practical use for it. Emulating consoles that have previously been based on PowerPC is moot as software emulators do an awesome job, and the mainstream gaming console platforms provide a way to play the the classic hits directly via their various backward compatibility features.

I suppose it could be useful as a card-based solution inside an x86-64 dev machine during ISA migration phase for AmigaOS/MorphOS.

Were more powerful and affordable FPGAs available earlier, it would change the nature of A-Eon's business by not relying on availability of PowerISA CPUs. Though they might have still worked with Varysis to produce a board for the A1-X series of computers. Then for certain the Xena slot would've been aimed at secondary FPGAs for 68k or ARM compatibility/development.

_________________
All the way, with 68k

@agami

Quote:

The market for and FPGA implementation of PowerPC cores is extremely small. Outside of idle curiosity, or the nostalgic desire to run AmigaOS 4.x there is no practical use for it. Emulating consoles that have previously been based on PowerPC is moot as software emulators do an awesome job, and the mainstream gaming console platforms provide a way to play the the classic hits directly via their various backward compatibility features.

18 months might get you something approaching similar performance level of an Efika so it wouldn't really be desirable.

Quote:

I suppose it could be useful as a card-based solution inside an x86-64 dev machine during ISA migration phase for AmigaOS/MorphOS.

I don't see how it would be helpful but it would be a huge drag on resources to make it work and major distraction from the end goal of migrating to a new ISA.

Instead...

-Put effort in getting FPGA development tools on AOS.

-Develop FPGA module that is usefull in classic Amigas.
--to get more I/O (GPIO+ethernet+USB etc.)
--to offload mpeg video + audio
--DCTV emulation + scandoubling
--on parallel port?

-Develop FPGA module that is usefull in NG devices.
--to get more I/O (GPIO+USB+8*joystickPort+USB-C etc.)
--mainstream needs 3GHz intel to read USB2.0 speed data (eg. Monsoon Inc PowerMonitor PSU), offload/buffer that
--custom USB that can do low CPU overhead USB + some special low latency operation (a little bit like xena)
--3D audio in+out processor
--glue logic to host ARM co-processor (ability to execute little endian binaries)
--HDMI capture (videotoaster -kind effects?)
--StarlinkInternet
--on PCIex1?

etc...

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@KimmoK

Quote:

-Develop FPGA module that is usefull in NG devices. --to get more I/O (GPIO+USB+8*joystickPort+USB-C etc.) --mainstream needs 3GHz intel to read USB2.0 speed data (eg. Monsoon Inc PowerMonitor PSU), offload/buffer that --custom USB that can do low CPU overhead USB + some special low latency operation (a little bit like xena) --3D audio in+out processor --glue logic to host ARM co-processor (ability to execute little endian binaries) --HDMI capture (videotoaster -kind effects?) --StarlinkInternet --on PCIex1? etc...

But not the obvious one - having Amiga chipsets available (and cycle exact 68000 if needed) available for the so called NG system?

Last edited by kolla on 11-Dec-2019 at 07:37 AM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Sure, why not also that.
+more https://drive.google.com/file/d/1fkOjp1uKgoZH9SKEeJTYiPrsnxHE0Zs-/view?usp=sharing



(some people would love to get authentic video output for old games also on NG box + to be able to get extra HDMI out + to be able to have real "Amiga" audio.)

With this kind of design, one could vampirize even SAM440ep-mini that has no PCIe:
https://voip.world/media/image/product/2129/md/allo-4pri-ec-pci-pcie-e1-t1-pri-2nd-gen-card-4-port-pci-pci-express-interfaces-on-the-same-board-echo-cancel_1.jpg

Last edited by KimmoK on 11-Dec-2019 at 11:26 AM.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

For now ppc hardware is good enough.
In the future, why not?
Amiga 68k with ppc in fpga, it looks interesting.

For prototyping OpenPower PowerPC core with some Amiga specific extra instructions, FPGA based powerpc would be useful. But to get >1ghz PPC performance we would need a real chip produced and I imagine that would cost million(s) of $. (cheaper to build sandbox + change ISA?)

+If PPC truly "dies", then also Risc-V with some extra PowerPC simulating instructions might become useful.

But so far it seems we have plenty of time to develop systems on PPC before availability ends.
(and there might appear more useful PowerPC/Power chips (that do not perform like A64, but that would be good enough) Today a4Ghz Power motherboard costs ~2000USD)


UPDATE:
If/once/when vampire chip is good enough and a real ~1Ghz chip could be made of it (preferably with a few PCIe lanes), then the need of other CPUs might vanish from our niche.

What is the future vision of Vampire development?
How huge bounty would be needed to get real new chip based on Vampire?

Last edited by KimmoK on 12-Dec-2019 at 09:35 AM.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

Vampire chip??

Vampire is a series of cards, not a chip.

Do you mean AC68080 in ASIC?

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@KimmoK
>How huge bounty would be needed to get real new chip based on Vampire?

@kolla
>Do you mean AC68080 in ASIC?

That's my assumption. And the cost there would certainly be more than a simple bounty could raise. Even just a basic ASIC design and run would cost upwards of a million USD or more; something like Vampire's 68080 I wouldn't be surprised to see a $10M or larger price tag attached.

@kolla

It would be most usefull as a system on chip.
It would then be useful in HW builds like RPi.

@coldacid

"I wouldn't be surprised to see a $10M or larger price tag attached."

Then we would need sponsoring from some party like NXP.

Last edited by KimmoK on 13-Dec-2019 at 06:46 AM.
Last edited by KimmoK on 13-Dec-2019 at 06:46 AM.
Last edited by KimmoK on 13-Dec-2019 at 06:45 AM.

_________________
- KimmoK
// For freedom, for honor, for AMIGA
//
// Thing that I should find more time for: CC64 - 64bit Community Computer?

@KimmoK

Why would anyone be interested in sponsoring a 68k ASIC?

@megol

Because it would be fully ASIC mate.

@Petah et al.,

It may have solved a few problems (and made others?) If A-eon put a FPGA with altivec and a compatible FPU instructions on the Tabor respin.

_________________
I sold my SAM460ex lite... waiting for money to buy a Raspberry Pi... or a Classic A1000 with Buffee... or an A1222... and OS4.3 FE update 11

@PhantomInterrogative

Only it just doesn't work that way....

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada